From 3668d2436480e5c1c78be8b2ce10f16cc59e390a Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Fri, 14 Aug 2026 19:04:43 +0000
Subject: [PATCH 01/22] SDK regeneration
---
.fern/metadata.json | 8 +-
.github/workflows/manual-publish.yml | 44 --
AGENTS.md | 101 ----
README.md | 49 +-
package.json | 26 +-
reference.md | 282 ++++++++-
src/BaseClient.ts | 6 +-
src/Client.ts | 6 +
src/api/errors/ContentTooLargeError.ts | 22 +
src/api/errors/index.ts | 1 +
src/api/resources/audio/client/Client.ts | 149 ++++-
.../audio/client/requests/GetSpeechRequest.ts | 16 +-
.../audio/client/requests/GetStreamRequest.ts | 60 --
.../client/requests/StreamAudioRequest.ts | 26 +
.../StreamWithTimestampsAudioRequest.ts | 24 +
.../resources/audio/client/requests/index.ts | 3 +-
.../StreamWithTimestampsAudioRequestAccept.ts | 10 +
src/api/resources/audio/types/index.ts | 1 +
src/api/resources/index.ts | 2 +
src/api/resources/models/client/Client.ts | 117 ++++
src/api/resources/models/client/index.ts | 1 +
src/api/resources/models/exports.ts | 4 +
src/api/resources/models/index.ts | 1 +
src/api/resources/voices/client/Client.ts | 63 +-
.../client/requests/CreateVoicesRequest.ts | 22 +-
.../client/requests/ListVoicesRequest.ts | 25 +-
src/api/resources/voices/exports.ts | 1 +
src/api/resources/voices/index.ts | 2 +
.../consentChallenges/client/Client.ts | 149 +++++
.../consentChallenges/client/index.ts | 1 +
.../requests/CreateConsentChallengeRequest.ts | 34 ++
.../client/requests/index.ts | 1 +
.../resources/consentChallenges/exports.ts | 4 +
.../resources/consentChallenges/index.ts | 1 +
src/api/resources/voices/resources/index.ts | 2 +
.../voices/types/ListVoicesRequestGender.ts | 8 +
.../voices/types/ListVoicesRequestType.ts | 7 +
src/api/resources/voices/types/index.ts | 2 +
src/api/types/ConsentChallenge.ts | 23 +
src/api/types/ErrorCode.ts | 18 +
src/api/types/ErrorDetail.ts | 7 +
src/api/types/Error_.ts | 4 +-
src/api/types/GetStreamRequest.ts | 46 ++
src/api/types/Model.ts | 66 ++
src/api/types/ModelsResponse.ts | 20 +
src/api/types/index.ts | 4 +
src/core/index.ts | 1 +
src/core/stream/Stream.ts | 235 ++++++++
src/core/stream/index.ts | 1 +
src/version.ts | 2 +-
tests/unit/stream/Stream.test.ts | 563 ++++++++++++++++++
tests/wire/audio.test.ts | 316 +++++++++-
tests/wire/models.test.ts | 183 ++++++
tests/wire/voices.test.ts | 5 +-
tests/wire/voices/consentChallenges.test.ts | 282 +++++++++
55 files changed, 2771 insertions(+), 286 deletions(-)
delete mode 100644 .github/workflows/manual-publish.yml
delete mode 100644 AGENTS.md
create mode 100644 src/api/errors/ContentTooLargeError.ts
delete mode 100644 src/api/resources/audio/client/requests/GetStreamRequest.ts
create mode 100644 src/api/resources/audio/client/requests/StreamAudioRequest.ts
create mode 100644 src/api/resources/audio/client/requests/StreamWithTimestampsAudioRequest.ts
create mode 100644 src/api/resources/audio/types/StreamWithTimestampsAudioRequestAccept.ts
create mode 100644 src/api/resources/models/client/Client.ts
create mode 100644 src/api/resources/models/client/index.ts
create mode 100644 src/api/resources/models/exports.ts
create mode 100644 src/api/resources/models/index.ts
create mode 100644 src/api/resources/voices/resources/consentChallenges/client/Client.ts
create mode 100644 src/api/resources/voices/resources/consentChallenges/client/index.ts
create mode 100644 src/api/resources/voices/resources/consentChallenges/client/requests/CreateConsentChallengeRequest.ts
create mode 100644 src/api/resources/voices/resources/consentChallenges/client/requests/index.ts
create mode 100644 src/api/resources/voices/resources/consentChallenges/exports.ts
create mode 100644 src/api/resources/voices/resources/consentChallenges/index.ts
create mode 100644 src/api/resources/voices/resources/index.ts
create mode 100644 src/api/resources/voices/types/ListVoicesRequestGender.ts
create mode 100644 src/api/resources/voices/types/ListVoicesRequestType.ts
create mode 100644 src/api/resources/voices/types/index.ts
create mode 100644 src/api/types/ConsentChallenge.ts
create mode 100644 src/api/types/GetStreamRequest.ts
create mode 100644 src/api/types/Model.ts
create mode 100644 src/api/types/ModelsResponse.ts
create mode 100644 src/core/stream/Stream.ts
create mode 100644 src/core/stream/index.ts
create mode 100644 tests/unit/stream/Stream.test.ts
create mode 100644 tests/wire/models.test.ts
create mode 100644 tests/wire/voices/consentChallenges.test.ts
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 78d1f36..7159737 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,9 +7,9 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "9f50f669dbb3d0db633a6d904db8a5077b94764f",
- "originGitCommitIsDirty": true,
+ "originGitCommit": "10a6db0bf8fe7358d20861af81ae01b0c177f5f0",
+ "originGitCommitIsDirty": false,
"invokedBy": "ci",
- "ciProvider": "unknown",
- "sdkVersion": "2.0.1"
+ "ciProvider": "github",
+ "sdkVersion": "3.0.2"
}
diff --git a/.github/workflows/manual-publish.yml b/.github/workflows/manual-publish.yml
deleted file mode 100644
index 61685b4..0000000
--- a/.github/workflows/manual-publish.yml
+++ /dev/null
@@ -1,44 +0,0 @@
-name: manual-publish
-
-# Manual, one-off publisher for an already-tagged release whose automatic
-# publish did not run (or failed). Checks out an explicit tag/ref, asserts
-# package.json version + repository.url, then publishes to npm with provenance.
-on:
- workflow_dispatch:
- inputs:
- ref:
- description: "Git tag/ref to publish (e.g. 3.0.0)"
- required: true
- default: "3.0.0"
- expected_version:
- description: "Version package.json MUST declare"
- required: true
- default: "3.0.0"
-
-jobs:
- publish:
- runs-on: ubuntu-latest
- permissions:
- contents: read
- id-token: write
- steps:
- - name: Checkout ref
- uses: actions/checkout@v4
- with:
- ref: ${{ inputs.ref }}
- - name: Set up node
- uses: actions/setup-node@v4
- with:
- node-version: "24"
- registry-url: "https://registry.npmjs.org"
- - name: Enable corepack
- run: corepack enable
- - name: Assert version + repository casing
- run: |
- node -e 'const p=require("./package.json"); if (p.version !== "${{ inputs.expected_version }}") throw new Error("version "+p.version); if (p.repository.url !== "git+https://github.com/SpeechifyInc/speechify-api-sdk-typescript.git") throw new Error("repository.url "+p.repository.url)'
- - name: Install dependencies
- run: pnpm install --frozen-lockfile
- - name: Build
- run: pnpm build
- - name: Publish to npm
- run: npm publish --provenance --access public
diff --git a/AGENTS.md b/AGENTS.md
deleted file mode 100644
index 11c5a9f..0000000
--- a/AGENTS.md
+++ /dev/null
@@ -1,101 +0,0 @@
-# Agent instructions — @speechify/api (TypeScript SDK)
-
-This is a **Fern-generated** SDK. It is published to **npm** as `@speechify/api`.
-The source of truth for the API surface lives in the `SpeechifyInc/speechify-api`
-repo (`fern/`); this repo receives generated code on the `sdk-release` branch.
-
-**Read this before doing anything that touches a release.** A botched release here
-publishes to a public registry and breaks the contract with every customer who
-installs the package. This file exists because we already did that once — see
-"Postmortem" below.
-
-## The golden rule
-
-**Publishing to npm is effectively IRREVERSIBLE.** Unpublish is restricted to a
-72-hour window and you can never republish the same version number. Treat every
-publish as permanent.
-
-Never trigger a publish until you have, in order:
-
-1. **Audited the release plumbing** (this file, the workflow, the config, npm auth).
-2. **Confirmed the version is correct in EVERY version-bearing file** (see checklist).
-3. **Dry-run built and asserted the artifact version** (`npm publish --dry-run`).
-4. **Got explicit human sign-off** for the publish itself.
-
-Do NOT admin-merge release PRs to force a publish. Do NOT bypass required reviews.
-
-## Version-surface checklist — the mistake we keep making
-
-The #1 failure mode is **a generated version string not getting bumped**, so the
-package publishes under the wrong number (or reports a false version to the API).
-A single stale string is a published contract break, not a typo.
-
-When any version changes, **confirm ALL of these agree** before release. Never
-check one and assume the rest:
-
-- `.release-please-manifest.json` → `"."`
-- `package.json` → `version`
-- `src/version.ts` → `SDK_VERSION` (sent as the SDK version header on requests)
-- the git tag created for the release
-- the version in the built tarball (`npm publish --dry-run` prints `version:`)
-
-Fast audit:
-
-```bash
-grep -rnE '"version"|SDK_VERSION|"\."' \
- .release-please-manifest.json package.json src/version.ts
-```
-
-`SDK_VERSION` is reported to the API. If it lies, telemetry, version-gating, and
-support debugging are all wrong for that release. (In the 3.0.0 release, the
-published tarball shipped `SDK_VERSION = "2.0.1"` — a live contract break.)
-
-## Known plumbing traps
-
-- **npm provenance requires case-exact `repository.url`.** `npm publish
- --provenance` fails with **HTTP 422** unless `package.json` `repository.url`
- matches the GitHub repo slug case-sensitively — it is `SpeechifyInc`, not
- `speechifyinc`. The casing is set by the generator config in
- `SpeechifyInc/speechify-api` (`fern/generators.yml`); fix it there so regens
- carry it.
-- **npm auth must exist for the publish to succeed.** The publish uses
- `--provenance` with `id-token: write`. That signs provenance via OIDC but does
- **NOT** by itself authenticate the upload — you still need EITHER a configured
- **npm Trusted Publisher (OIDC)** for `@speechify/api` linked to this repo +
- workflow, OR an `NPM_TOKEN` secret wired as `NODE_AUTH_TOKEN`. Without one, the
- publish fails with **HTTP 404 "no permission"**. Verify auth BEFORE relying on
- the automatic publish.
-- **`type: "generic"` extra-files no-op silently** unless the target line carries
- an `x-release-please-version` marker comment. `src/version.ts` is bumped this
- way — confirm it actually changed after a release, or fix it manually.
-- **npm account 2FA is a passkey.** The CLI `--otp=` flow only accepts TOTP, not
- passkeys, so a terminal `npm publish` prompts `EOTP` and cannot proceed with a
- passkey. Use `npm login --auth-type=web` (browser passkey) for a local publish,
- or an automation token / Trusted Publisher for CI.
-
-## If a release has already gone wrong
-
-- **Wrong version on npm:** treat as permanent. Fix the version wiring, cut a NEW
- correct version. Do not count on unpublish.
-- **Tag points at a bad commit:** recreate the tag/GitHub release on the corrected
- commit (public history mutation — back up the old SHA + notes first). Publishing
- from a stale tag re-ships the old bug (e.g. the lowercase `repository.url`).
-- There is a `manual-publish.yml` workflow that checks out an explicit tag and
- asserts `version` + `repository.url` before publishing — prefer it for one-off
- republishes.
-
-## Postmortem — the 3.0.0 release (why this file exists)
-
-A routine regeneration was pushed straight to publish. Failures, in order:
-
-1. Breaking-change PRs were admin-merged past the review gate.
-2. The automatic publish failed provenance (HTTP 422) because `repository.url` was
- lowercase `speechifyinc` — npm was left at 2.0.0 while the tag said 3.0.0.
-3. After fixing casing, publish failed again (HTTP 404) — no npm auth was
- configured (no `NPM_TOKEN`, no Trusted Publisher).
-4. When it finally published, the tarball still shipped `SDK_VERSION = "2.0.1"`
- because the generic updater had silently no-oped — found reactively, late.
-
-Lesson: **audit the whole release surface first (versions AND auth AND
-provenance), dry-run, get sign-off, then publish.** Treat one stale version
-string as a signal to check every other one.
diff --git a/README.md b/README.md
index 00c54a4..92bfda0 100644
--- a/README.md
+++ b/README.md
@@ -1,6 +1,6 @@
# Speechify TypeScript Library
-[](https://buildwithfern.com?utm_source=github&utm_medium=github&utm_campaign=readme&utm_source=https%3A%2F%2Fgithub.com%2Fspeechifyinc%2Fspeechify-api-sdk-typescript)
+[](https://buildwithfern.com?utm_source=github&utm_medium=github&utm_campaign=readme&utm_source=https%3A%2F%2Fgithub.com%2Fspeechify-ai%2Fsdk-typescript)
[](https://www.npmjs.com/package/@speechify/api)
The Speechify TypeScript library provides convenient access to the Speechify APIs from TypeScript.
@@ -14,6 +14,7 @@ The Speechify TypeScript library provides convenient access to the Speechify API
- [Environments](#environments)
- [Request and Response Types](#request-and-response-types)
- [Exception Handling](#exception-handling)
+- [Streaming Response](#streaming-response)
- [File Uploads](#file-uploads)
- [Binary Response](#binary-response)
- [Pagination](#pagination)
@@ -42,7 +43,7 @@ npm i -s @speechify/api
## Reference
-A full reference for this library is available [here](https://github.com/speechifyinc/speechify-api-sdk-typescript/blob/HEAD/./reference.md).
+A full reference for this library is available [here](https://github.com/speechify-ai/sdk-typescript/blob/HEAD/./reference.md).
## Usage
@@ -51,12 +52,12 @@ Instantiate and use the client with the following:
```typescript
import { SpeechifyClient } from "@speechify/api";
-const client = new SpeechifyClient({ token: "YOUR_TOKEN", version: "2026-07-07" });
+const client = new SpeechifyClient({ token: "YOUR_TOKEN", version: "2026-09-13" });
await client.audio.speech({
audio_format: "mp3",
input: "Hello! This is the Speechify text-to-speech API.",
- model: "simba-english",
- voice_id: "george"
+ model: "simba-3.2",
+ voice_id: "geffen_32"
});
```
@@ -105,6 +106,27 @@ try {
}
```
+## Streaming Response
+
+Some endpoints return streaming responses instead of returning the full response at once.
+The SDK uses async iterators, so you can consume the responses using a `for await...of` loop.
+
+```typescript
+import { SpeechifyClient } from "@speechify/api";
+
+const client = new SpeechifyClient({ token: "YOUR_TOKEN", version: "2026-09-13" });
+const response = await client.audio.streamWithTimestamps({
+ body: {
+ input: "Streaming long-form audio with the Speechify API.",
+ model: "simba-3.2",
+ voice_id: "geffen_32"
+ }
+});
+for await (const item of response) {
+ console.log(item);
+}
+```
+
## File Uploads
You can upload files using the client:
@@ -114,13 +136,14 @@ import { createReadStream } from "fs";
import * as fs from "fs";
import { SpeechifyClient } from "@speechify/api";
-const client = new SpeechifyClient({ token: "YOUR_TOKEN", version: "2026-07-07" });
+const client = new SpeechifyClient({ token: "YOUR_TOKEN", version: "2026-09-13" });
await client.voices.create({
sample: fs.createReadStream("/path/to/your/file"),
+ consent_recording: fs.createReadStream("/path/to/your/file"),
"Idempotency-Key": "a1b2c3d4-5e6f-7a8b-9c0d-1e2f3a4b5c6d",
name: "name",
gender: "male",
- consent: "consent"
+ consent_challenge_id: "consent_challenge_id"
});
```
The client accepts a variety of types for file upload parameters:
@@ -549,14 +572,20 @@ List endpoints are paginated. The SDK provides an iterator so that you can simpl
```typescript
import { SpeechifyClient } from "@speechify/api";
-const client = new SpeechifyClient({ token: "YOUR_TOKEN", version: "2026-07-07" });
-const pageableResponse = await client.voices.list();
+const client = new SpeechifyClient({ token: "YOUR_TOKEN", version: "2026-09-13" });
+const pageableResponse = await client.voices.list({
+ locale: "en",
+ model: "simba-3.2"
+});
for await (const item of pageableResponse) {
console.log(item);
}
// Or you can manually iterate page-by-page
-let page = await client.voices.list();
+let page = await client.voices.list({
+ locale: "en",
+ model: "simba-3.2"
+});
while (page.hasNextPage()) {
page = page.getNextPage();
}
diff --git a/package.json b/package.json
index 2ef7387..7b12553 100644
--- a/package.json
+++ b/package.json
@@ -1,10 +1,10 @@
{
"name": "@speechify/api",
- "version": "3.0.1",
+ "version": "3.0.2",
"private": false,
"repository": {
"type": "git",
- "url": "git+https://github.com/SpeechifyInc/speechify-api-sdk-typescript.git"
+ "url": "git+https://github.com/speechify-ai/sdk-typescript.git"
},
"type": "commonjs",
"main": "./dist/cjs/index.js",
@@ -33,6 +33,17 @@
},
"default": "./dist/cjs/api/resources/audio/exports.js"
},
+ "./models": {
+ "import": {
+ "types": "./dist/esm/api/resources/models/exports.d.mts",
+ "default": "./dist/esm/api/resources/models/exports.mjs"
+ },
+ "require": {
+ "types": "./dist/cjs/api/resources/models/exports.d.ts",
+ "default": "./dist/cjs/api/resources/models/exports.js"
+ },
+ "default": "./dist/cjs/api/resources/models/exports.js"
+ },
"./voices": {
"import": {
"types": "./dist/esm/api/resources/voices/exports.d.mts",
@@ -44,6 +55,17 @@
},
"default": "./dist/cjs/api/resources/voices/exports.js"
},
+ "./voices/consentChallenges": {
+ "import": {
+ "types": "./dist/esm/api/resources/voices/resources/consentChallenges/exports.d.mts",
+ "default": "./dist/esm/api/resources/voices/resources/consentChallenges/exports.mjs"
+ },
+ "require": {
+ "types": "./dist/cjs/api/resources/voices/resources/consentChallenges/exports.d.ts",
+ "default": "./dist/cjs/api/resources/voices/resources/consentChallenges/exports.js"
+ },
+ "default": "./dist/cjs/api/resources/voices/resources/consentChallenges/exports.js"
+ },
"./package.json": "./package.json"
},
"files": [
diff --git a/reference.md b/reference.md
index 8ff4747..28c6a5b 100644
--- a/reference.md
+++ b/reference.md
@@ -34,8 +34,8 @@ Set `output_format` for explicit sample-rate/bitrate control (e.g.
await client.audio.speech({
audio_format: "mp3",
input: "Hello! This is the Speechify text-to-speech API.",
- model: "simba-english",
- voice_id: "george"
+ model: "simba-3.2",
+ voice_id: "geffen_32"
});
```
@@ -106,8 +106,10 @@ POST /v1/audio/speech.
```typescript
await client.audio.stream({
- input: "input",
- voice_id: "voice_id"
+ body: {
+ input: "input",
+ voice_id: "voice_id"
+ }
});
```
@@ -124,7 +126,7 @@ await client.audio.stream({
-
-**request:** `Speechify.GetStreamRequest`
+**request:** `Speechify.StreamAudioRequest`
@@ -140,6 +142,166 @@ await client.audio.stream({
+
+
+
+
+client.audio.streamWithTimestamps({ ...params }) -> core.Stream<string>
+
+-
+
+#### 📝 Description
+
+
+-
+
+
+-
+
+Synthesize speech and stream it back together with word-level speech
+marks, for text highlighting, captions and audio-text synchronization
+while the audio is still arriving.
+
+The response is a Server-Sent Events stream. Each `speech.chunk` event
+carries a Base64-encoded run of audio, the speech marks that became
+final with it, or both - a chunk may carry only one of the two, and the
+last chunk of a stream is often marks-only. A terminal `speech.done`
+event ends the stream; there is no `[DONE]` sentinel. Ignore any event
+type you do not recognize, so that new event types do not break your
+integration.
+
+Speech-mark times are absolute milliseconds from the start of the
+synthesis, so concatenate the audio chunks into one stream and apply the
+marks against that single timeline. Which chunk a mark arrives on is a
+delivery detail and carries no meaning. Times stay correct for every
+`output_format`: changing the codec or sample rate does not change the
+duration.
+
+Speech marks are produced by the streaming-native models. The default
+`simba-3.0` and `simba-3.2` both serve this route; the legacy
+`simba-english` and `simba-multilingual` models return 400
+`speech_marks_unsupported` here.
+For Base64-encoded audio and speech marks in one non-streamed JSON
+response, on any model, use POST /v1/audio/speech.
+
+
+
+
+
+#### 🔌 Usage
+
+
+-
+
+
+-
+
+```typescript
+const response = await client.audio.streamWithTimestamps({
+ body: {
+ input: "Streaming long-form audio with the Speechify API.",
+ model: "simba-3.2",
+ voice_id: "geffen_32"
+ }
+});
+for await (const item of response) {
+ console.log(item);
+}
+
+```
+
+
+
+
+
+#### ⚙️ Parameters
+
+
+-
+
+
+-
+
+**request:** `Speechify.StreamWithTimestampsAudioRequest`
+
+
+
+
+
+-
+
+**requestOptions:** `AudioClient.RequestOptions`
+
+
+
+
+
+
+
+
+
+client.models.list() -> Speechify.ModelsResponse
+
+-
+
+#### 📝 Description
+
+
+-
+
+
+-
+
+List the text-to-speech models available for synthesis. Drive a model
+picker from this response, then pass a model `id` as the `model`
+parameter to POST /v1/audio/speech or /v1/audio/stream. The response
+marks the default model (used when a request omits `model`), the
+routes each model may be passed to, and which voices it accepts.
+Multi-speaker models arrive in a separate `dialogue_models` array
+because they are valid only on POST /v1/audio/dialogue. Returns
+the full set in a single response: the model catalog is static
+platform reference data, so it is intentionally not paginated.
+
+
+
+
+
+#### 🔌 Usage
+
+
+-
+
+
+-
+
+```typescript
+await client.models.list();
+
+```
+
+
+
+
+
+#### ⚙️ Parameters
+
+
+-
+
+
+-
+
+**requestOptions:** `ModelsClient.RequestOptions`
+
+
+
+
+
+
+
-
-Create a personal (cloned) voice for the user
+Create a cloned voice for the workspace from a 10-30 second audio sample, with verified consent from the speaker.
+
+Cloning requires proof that the speaker agreed to it. Create a consent challenge with `POST /v1/voices/consent-challenges`, show the returned `phrase` to the speaker, record them reading it aloud, and send that recording here as `consent_recording` together with the challenge's `consent_challenge_id`. Speechify transcribes the recording, checks it against the phrase it issued, and keeps it as the consent record for the voice. A challenge is single use and short-lived, so record and submit in one sitting.
+
+The clone belongs to the workspace rather than the member who created it, and access follows the caller's workspace role and API-key scopes exactly as for any other voice: voices scopes to list it, audio scopes to synthesize with it, and the content-management permission plus a write scope on the key to delete it. Cloned voices are usable self-serve on `simba-3.0`, `simba-english` and `simba-multilingual`. `simba-3.2` also serves cloned voices, currently as a limited release enabled per workspace; contact Speechify to have it enabled for yours.
+
+Callers pinned before `Speechify-Version: 2026-09-13` use the previous flow instead: no challenge, and a `consent` form field carrying the speaker's name and email as a JSON string. That flow is deprecated and will be removed after a sunset window announced in the changelog.
@@ -254,10 +430,11 @@ Create a personal (cloned) voice for the user
```typescript
await client.voices.create({
sample: fs.createReadStream("/path/to/your/file"),
+ consent_recording: fs.createReadStream("/path/to/your/file"),
"Idempotency-Key": "a1b2c3d4-5e6f-7a8b-9c0d-1e2f3a4b5c6d",
name: "name",
gender: "male",
- consent: "consent"
+ consent_challenge_id: "consent_challenge_id"
});
```
@@ -307,9 +484,9 @@ await client.voices.create({
Fetch a single voice by id - a shared catalogue voice or one of
-the caller's own personal (cloned) voices. A personal voice that
-belongs to another workspace returns 404, identical to an
-unknown id, so voice inventory is never enumerable across tenants.
+the workspace's cloned voices. A cloned voice that belongs to
+another workspace returns 404, identical to an unknown id, so
+voice inventory is never enumerable across tenants.
@@ -374,7 +551,9 @@ await client.voices.get({
-
-Delete a personal (cloned) voice
+Delete one of the workspace's cloned voices. Requires the
+`content.manage` permission (owner, admin, or member); a
+service-account key is authorized by its scopes instead.
@@ -492,3 +671,76 @@ await client.voices.downloadSample({
+## Voices ConsentChallenges
+client.voices.consentChallenges.create({ ...params }) -> Speechify.ConsentChallenge
+
+-
+
+#### 📝 Description
+
+
+-
+
+
+-
+
+Start the consent check for a voice clone.
+
+Returns a `phrase` for the speaker to read aloud and an `id` that identifies this challenge. Show the phrase to the speaker exactly as returned, record them reading it, then send the recording and the `id` to `POST /v1/voices`, which verifies the recording against the phrase and keeps it as the consent record.
+
+A challenge is single use, is bound to the workspace that created it, and expires at `expires_at` - it is proof that a speaker was in front of a microphone just now, so create it when you are ready to record, not at the start of your flow. If it expires, create another one and record again.
+
+Challenge creation is rate limited per workspace at a few dozen per hour, far more tightly than the rest of the voice surface, because each one precedes a person recording themselves - mint it when your speaker is ready, not speculatively. Read the live ceiling off `RateLimit-*` rather than hard-coding it. **On a `429`, always honour `Retry-After` rather than a fixed backoff of your own**: the wait is measured in minutes and can run to most of an hour. `RateLimit-*` are omitted rather than reporting a bucket that is not the one refusing.
+
+
+
+
+
+#### 🔌 Usage
+
+
+-
+
+
+-
+
+```typescript
+await client.voices.consentChallenges.create({
+ "Idempotency-Key": "a1b2c3d4-5e6f-7a8b-9c0d-1e2f3a4b5c6d",
+ full_name: "Jane Doe"
+});
+
+```
+
+
+
+
+
+#### ⚙️ Parameters
+
+
+-
+
+
+-
+
+**request:** `Speechify.voices.CreateConsentChallengeRequest`
+
+
+
+
+
+-
+
+**requestOptions:** `ConsentChallengesClient.RequestOptions`
+
+
+
+
+
+
+
+
+
+client.audio.streamWithTimestamps({ ...params }) -> core.Stream<string>
+client.audio.streamWithTimestamps({ ...params }) -> core.Stream<Speechify.SpeechStreamEvent>
-
diff --git a/src/api/resources/audio/client/Client.ts b/src/api/resources/audio/client/Client.ts
index 053283c..c041972 100644
--- a/src/api/resources/audio/client/Client.ts
+++ b/src/api/resources/audio/client/Client.ts
@@ -287,14 +287,14 @@ export class AudioClient {
public streamWithTimestamps(
request: Speechify.StreamWithTimestampsAudioRequest,
requestOptions?: AudioClient.RequestOptions,
- ): core.HttpResponsePromise> {
+ ): core.HttpResponsePromise> {
return core.HttpResponsePromise.fromPromise(this.__streamWithTimestamps(request, requestOptions));
}
private async __streamWithTimestamps(
request: Speechify.StreamWithTimestampsAudioRequest,
requestOptions?: AudioClient.RequestOptions,
- ): Promise>> {
+ ): Promise>> {
const { Accept: accept, body: _body } = request;
const _authRequest: core.AuthRequest = await this._options.authProvider.getAuthRequest();
const _headers: core.Fetcher.Args["headers"] = mergeHeaders(
diff --git a/src/api/types/SpeechChunkEvent.ts b/src/api/types/SpeechChunkEvent.ts
new file mode 100644
index 0000000..e02900d
--- /dev/null
+++ b/src/api/types/SpeechChunkEvent.ts
@@ -0,0 +1,26 @@
+// This file was auto-generated by Fern from our API Definition.
+
+import type * as Speechify from "../index.js";
+
+/**
+ * A run of synthesized audio, the speech marks that became final with it,
+ * or both - a chunk may carry only one of the two, and the last chunk of
+ * a stream is often marks-only. Mark times are absolute milliseconds from
+ * the start of the synthesis: concatenate the audio chunks into one
+ * stream and apply the marks against that single timeline. Which chunk a
+ * mark arrives on is a delivery detail and carries no meaning.
+ */
+export interface SpeechChunkEvent {
+ /**
+ * A run of the synthesized audio, Base64-encoded, in the format the
+ * request selected (echoed on the `Speechify-Audio-Content-Type`
+ * response header). Absent on a marks-only chunk.
+ */
+ audio?: string | undefined;
+ /**
+ * Word timings addressing the original input text, with absolute
+ * millisecond times from the start of the synthesis. Absent when the
+ * chunk carries only audio.
+ */
+ speech_marks?: Speechify.NestedChunk[] | undefined;
+}
diff --git a/src/api/types/SpeechDoneEvent.ts b/src/api/types/SpeechDoneEvent.ts
new file mode 100644
index 0000000..174ce8b
--- /dev/null
+++ b/src/api/types/SpeechDoneEvent.ts
@@ -0,0 +1,12 @@
+// This file was auto-generated by Fern from our API Definition.
+
+/**
+ * Terminal event: the synthesis completed and no further events follow.
+ * There is no `[DONE]` sentinel.
+ */
+export interface SpeechDoneEvent {
+ /** Number of billable characters processed. */
+ billable_characters_count: number;
+ /** Duration of the synthesized audio in milliseconds. */
+ audio_duration_ms: number;
+}
diff --git a/src/api/types/SpeechErrorEvent.ts b/src/api/types/SpeechErrorEvent.ts
new file mode 100644
index 0000000..3109b98
--- /dev/null
+++ b/src/api/types/SpeechErrorEvent.ts
@@ -0,0 +1,17 @@
+// This file was auto-generated by Fern from our API Definition.
+
+import type * as Speechify from "../index.js";
+
+/**
+ * Terminal event carrying the standard error envelope, emitted when a
+ * failure happens after the stream has started and the status code is
+ * already committed.
+ */
+export interface SpeechErrorEvent {
+ error: Speechify.ErrorDetail;
+ /**
+ * Server-side request identifier. Echoes the `Speechify-Request-Id`
+ * response header.
+ */
+ request_id?: string | undefined;
+}
diff --git a/src/api/types/SpeechStreamEvent.ts b/src/api/types/SpeechStreamEvent.ts
new file mode 100644
index 0000000..165581f
--- /dev/null
+++ b/src/api/types/SpeechStreamEvent.ts
@@ -0,0 +1,27 @@
+// This file was auto-generated by Fern from our API Definition.
+
+import type * as Speechify from "../index.js";
+
+/**
+ * One event on the POST /v1/audio/stream/with-timestamps stream. The
+ * `type` field discriminates the variants and mirrors the SSE `event:`
+ * name, so an event is identifiable from its `data:` payload alone.
+ */
+export type SpeechStreamEvent =
+ | Speechify.SpeechStreamEvent.SpeechChunk
+ | Speechify.SpeechStreamEvent.SpeechDone
+ | Speechify.SpeechStreamEvent.SpeechError;
+
+export namespace SpeechStreamEvent {
+ export interface SpeechChunk extends Speechify.SpeechChunkEvent {
+ type: "speech.chunk";
+ }
+
+ export interface SpeechDone extends Speechify.SpeechDoneEvent {
+ type: "speech.done";
+ }
+
+ export interface SpeechError extends Speechify.SpeechErrorEvent {
+ type: "speech.error";
+ }
+}
diff --git a/src/api/types/index.ts b/src/api/types/index.ts
index be37ea1..6793d96 100644
--- a/src/api/types/index.ts
+++ b/src/api/types/index.ts
@@ -16,4 +16,8 @@ export * from "./Model.js";
export * from "./ModelsResponse.js";
export * from "./NestedChunk.js";
export * from "./PaginationMeta.js";
+export * from "./SpeechChunkEvent.js";
+export * from "./SpeechDoneEvent.js";
+export * from "./SpeechErrorEvent.js";
export * from "./SpeechMarks.js";
+export * from "./SpeechStreamEvent.js";
diff --git a/tests/wire/audio.test.ts b/tests/wire/audio.test.ts
index 8d7e96c..132e360 100644
--- a/tests/wire/audio.test.ts
+++ b/tests/wire/audio.test.ts
@@ -318,7 +318,8 @@ describe("AudioClient", () => {
model: "simba-3.2",
voice_id: "geffen_32",
};
- const rawResponseBody = 'event: \ndata: "string"\n\n';
+ const rawResponseBody =
+ 'event: \ndata: {"audio":"SUQzBAAAAAAAI1RTU0UAAAAPAAADTGF2ZjYxLjcuMTAw...","speech_marks":[{"end":5,"end_time":320,"start":0,"start_time":0,"type":"word","value":"Hello"}],"type":"speech.chunk"}\n\n';
server
.mockEndpoint()
@@ -340,7 +341,22 @@ describe("AudioClient", () => {
for await (const event of response) {
events.push(event);
}
- expect(events).toEqual(["string"]);
+ expect(events).toEqual([
+ {
+ type: "speech.chunk",
+ audio: "SUQzBAAAAAAAI1RTU0UAAAAPAAADTGF2ZjYxLjcuMTAw...",
+ speech_marks: [
+ {
+ end: 5,
+ end_time: 320,
+ start: 0,
+ start_time: 0,
+ type: "word",
+ value: "Hello",
+ },
+ ],
+ },
+ ]);
});
test("streamWithTimestamps (2)", async () => {
From 815185f26d03d4e51c7ca17af02863edb0942cca Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Mon, 17 Aug 2026 18:17:55 +0000
Subject: [PATCH 08/22] SDK regeneration
---
.fern/metadata.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index cf3135a..856661a 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "d33e7cfed6bc7c034884ec3ccf898f742cf1abf4",
+ "originGitCommit": "181fb6e145a080cb2cb044fa64a7c5e2638bd66d",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
From 596ea1ed065b0582ec8192e2c887c46d42936a49 Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Mon, 17 Aug 2026 20:46:50 +0000
Subject: [PATCH 09/22] SDK regeneration
---
.fern/metadata.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 856661a..7a15477 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "181fb6e145a080cb2cb044fa64a7c5e2638bd66d",
+ "originGitCommit": "9b87dd733251b3bb18e005106fbe7c1ba651bc7a",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
From e1bafd2514eda040741e5fc1414417d2d9be7cc2 Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Mon, 17 Aug 2026 20:49:19 +0000
Subject: [PATCH 10/22] SDK regeneration
From 4212218f1a09ad2d1d8e78858d464843827e3fb2 Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Tue, 18 Aug 2026 08:51:30 +0000
Subject: [PATCH 11/22] SDK regeneration
---
.fern/metadata.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 7a15477..235620f 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "9b87dd733251b3bb18e005106fbe7c1ba651bc7a",
+ "originGitCommit": "03f284ab3f90b26e4bb7a82c8aa9ac7573d2d7e0",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
From 618663af95ae9ffe3c340c171dab140110c6c49b Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Tue, 18 Aug 2026 10:21:35 +0000
Subject: [PATCH 12/22] SDK regeneration
---
.fern/metadata.json | 2 +-
tests/wire/models.test.ts | 4 ++--
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 235620f..f6979c0 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "03f284ab3f90b26e4bb7a82c8aa9ac7573d2d7e0",
+ "originGitCommit": "950720e1d28678ca996c35f429b566b81aeef303",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
diff --git a/tests/wire/models.test.ts b/tests/wire/models.test.ts
index 00efd8d..931027b 100644
--- a/tests/wire/models.test.ts
+++ b/tests/wire/models.test.ts
@@ -18,7 +18,7 @@ describe("ModelsClient", () => {
models: [
{
id: "simba-english",
- name: "Simba English",
+ name: "Simba 1.6 English",
default: false,
recommended: false,
deprecated: true,
@@ -30,7 +30,7 @@ describe("ModelsClient", () => {
},
{
id: "simba-multilingual",
- name: "Simba Multilingual",
+ name: "Simba 1.6 Multilingual",
default: false,
recommended: false,
deprecated: true,
From b5db64f2416c75acf4873473f706514c6aea151f Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Tue, 18 Aug 2026 11:09:43 +0000
Subject: [PATCH 13/22] SDK regeneration
---
.fern/metadata.json | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index f6979c0..b49b33c 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "950720e1d28678ca996c35f429b566b81aeef303",
+ "originGitCommit": "70772af113ee9325533287ba25c90589a4eb99b8",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
From 56de1777df31929ce00d85cdf45d2ed2ab8c790a Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Tue, 18 Aug 2026 11:37:28 +0000
Subject: [PATCH 14/22] SDK regeneration
---
.fern/metadata.json | 2 +-
src/api/types/AudioOutputFormat.ts | 8 +++++++-
src/api/types/AudioStreamOutputFormat.ts | 8 +++++++-
src/api/types/GetSpeechResponse.ts | 2 +-
4 files changed, 16 insertions(+), 4 deletions(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index b49b33c..54cb320 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "70772af113ee9325533287ba25c90589a4eb99b8",
+ "originGitCommit": "87d84f2414b409b5f97f46bd95b076864ce9d39f",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
diff --git a/src/api/types/AudioOutputFormat.ts b/src/api/types/AudioOutputFormat.ts
index b24502e..f1c7c5c 100644
--- a/src/api/types/AudioOutputFormat.ts
+++ b/src/api/types/AudioOutputFormat.ts
@@ -1,6 +1,10 @@
// This file was auto-generated by Fern from our API Definition.
-/** Audio output format as a `codec_sampleRate_bitrate` string, giving explicit control over sample rate and bitrate. `pcm_*` and `ulaw_8000` are headerless raw audio; `pcm_16000` and `ulaw_8000` are the telephony formats Twilio/LiveKit SIP expect. */
+/**
+ * Audio output format as a `codec_sampleRate_bitrate` string, giving explicit control over sample rate and bitrate. `pcm_*` and `ulaw_8000` are headerless raw audio; `pcm_16000` and `ulaw_8000` are the telephony formats Twilio/LiveKit SIP expect.
+ *
+ * 160 kbps is the highest bitrate an mp3 can carry at 22.05 and 24 kHz, so `mp3_22050_160` and `mp3_24000_160` are the maximum-fidelity mp3 formats; a request for `mp3_*_192` is encoded at 160 kbps and reported as the matching `mp3_*_160`. The two `mp3_*_160` formats are served by the Simba 3 models only.
+ */
export const AudioOutputFormat = {
Pcm8000: "pcm_8000",
Pcm16000: "pcm_16000",
@@ -12,11 +16,13 @@ export const AudioOutputFormat = {
Mp32205064: "mp3_22050_64",
Mp32205096: "mp3_22050_96",
Mp322050128: "mp3_22050_128",
+ Mp322050160: "mp3_22050_160",
Mp322050192: "mp3_22050_192",
Mp32400032: "mp3_24000_32",
Mp32400064: "mp3_24000_64",
Mp32400096: "mp3_24000_96",
Mp324000128: "mp3_24000_128",
+ Mp324000160: "mp3_24000_160",
Mp324000192: "mp3_24000_192",
Wav24000: "wav_24000",
Wav48000: "wav_48000",
diff --git a/src/api/types/AudioStreamOutputFormat.ts b/src/api/types/AudioStreamOutputFormat.ts
index 0013c28..fa16e83 100644
--- a/src/api/types/AudioStreamOutputFormat.ts
+++ b/src/api/types/AudioStreamOutputFormat.ts
@@ -1,6 +1,10 @@
// This file was auto-generated by Fern from our API Definition.
-/** Audio output format for the streaming endpoint (`POST /v1/audio/stream`), as a `codec_sampleRate_bitrate` string. Same as `AudioOutputFormat` minus the `wav_*` formats: wav is only available on `POST /v1/audio/speech`. `pcm_*` and `ulaw_8000` are headerless raw audio; `pcm_16000` and `ulaw_8000` are the telephony formats Twilio/LiveKit SIP expect. */
+/**
+ * Audio output format for the streaming endpoint (`POST /v1/audio/stream`), as a `codec_sampleRate_bitrate` string. Same as `AudioOutputFormat` minus the `wav_*` formats: wav is only available on `POST /v1/audio/speech`. `pcm_*` and `ulaw_8000` are headerless raw audio; `pcm_16000` and `ulaw_8000` are the telephony formats Twilio/LiveKit SIP expect.
+ *
+ * 160 kbps is the highest bitrate an mp3 can carry at 22.05 and 24 kHz, so `mp3_22050_160` and `mp3_24000_160` are the maximum-fidelity mp3 formats; a request for `mp3_*_192` is encoded at 160 kbps. The two `mp3_*_160` formats are served by the Simba 3 models only.
+ */
export const AudioStreamOutputFormat = {
Pcm8000: "pcm_8000",
Pcm16000: "pcm_16000",
@@ -12,11 +16,13 @@ export const AudioStreamOutputFormat = {
Mp32205064: "mp3_22050_64",
Mp32205096: "mp3_22050_96",
Mp322050128: "mp3_22050_128",
+ Mp322050160: "mp3_22050_160",
Mp322050192: "mp3_22050_192",
Mp32400032: "mp3_24000_32",
Mp32400064: "mp3_24000_64",
Mp32400096: "mp3_24000_96",
Mp324000128: "mp3_24000_128",
+ Mp324000160: "mp3_24000_160",
Mp324000192: "mp3_24000_192",
Ulaw8000: "ulaw_8000",
Ogg24000: "ogg_24000",
diff --git a/src/api/types/GetSpeechResponse.ts b/src/api/types/GetSpeechResponse.ts
index 83cada3..b4b58c9 100644
--- a/src/api/types/GetSpeechResponse.ts
+++ b/src/api/types/GetSpeechResponse.ts
@@ -9,7 +9,7 @@ export interface GetSpeechResponse {
audio_format: GetSpeechResponse.AudioFormat;
/** The number of billable characters processed in the request. */
billable_characters_count: number;
- /** The full `codec_sampleRate_bitrate` format, echoed back when the request set `output_format`. */
+ /** The full `codec_sampleRate_bitrate` format the audio was encoded in, returned when the request set `output_format`. It is the requested value unless the request named a bitrate above the mp3 ceiling, in which case it reports the bitrate actually delivered. */
output_format?: Speechify.AudioOutputFormat | undefined;
speech_marks: Speechify.SpeechMarks;
}
From c6943308e855b4b7dfae96af2afc007142fe5bb5 Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Tue, 18 Aug 2026 15:20:04 +0000
Subject: [PATCH 15/22] SDK regeneration
---
.fern/metadata.json | 8 ++++----
package.json | 2 +-
src/BaseClient.ts | 4 ++--
src/version.ts | 2 +-
4 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 54cb320..5821670 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,9 +7,9 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "87d84f2414b409b5f97f46bd95b076864ce9d39f",
+ "originGitCommit": "469722e153faa1a1c50aa63ea9cdca853f588721",
"originGitCommitIsDirty": false,
- "invokedBy": "ci",
- "ciProvider": "github",
- "sdkVersion": "3.0.2"
+ "invokedBy": "manual",
+ "requestedVersion": "4.0.0",
+ "sdkVersion": "4.0.0"
}
diff --git a/package.json b/package.json
index 0ab9286..beff59d 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
{
"name": "@speechify/api",
- "version": "3.0.2",
+ "version": "4.0.0",
"private": false,
"repository": {
"type": "git",
diff --git a/src/BaseClient.ts b/src/BaseClient.ts
index 2d206f1..56fc834 100644
--- a/src/BaseClient.ts
+++ b/src/BaseClient.ts
@@ -63,8 +63,8 @@ export function normalizeClientOptions
Date: Tue, 18 Aug 2026 19:03:12 +0100
Subject: [PATCH 16/22] fix: discriminate SSE events on type and harden stream
abort
streamWithTimestamps constructed core.Stream with no eventDiscriminator,
routing SSE through iterDataMessages(). That parser drops the event: name,
cannot concatenate a multi-line data: field, and discards a trailing event
with no terminating blank line. The discarded trailing event is speech.done,
which carries billable_characters_count. Passing eventDiscriminator: "type"
routes to iterSseEvents() and fixes all three.
The live API sends a populated event: speech.chunk name and also duplicates
type in the payload. The generated wire fixture showing an empty event: does
not match the API.
core.Stream assigned a private AbortController that nothing read, and
registered an abort listener closing over this. The handler was therefore a
no-op, and a reused AbortSignal retained every Stream. Both are replaced with
throwIfAborted().
The abort unit test passed on its own break statement rather than on the
abort. It is rewritten so it fails when abort breaks.
README and reference streaming examples now narrow on item.type, handle
speech.error (yielded, never thrown, because the 200 is already committed)
and show the base64 decode.
.fernignore shields these until the Fern TypeScript generator emits the
discriminator. fern-python-sdk@5.14.20 already generates a correct
discriminated union on type from the same API definition;
fern-typescript-sdk@3.70.1 emits none.
---
.fernignore | 49 +++++++
README.md | 37 ++++-
reference.md | 37 ++++-
src/api/resources/audio/client/Client.ts | 1 +
src/core/stream/Stream.ts | 27 +++-
tests/unit/stream/Stream.test.ts | 176 +++++++++++++++++++++--
6 files changed, 314 insertions(+), 13 deletions(-)
diff --git a/.fernignore b/.fernignore
index 0ae32ea..4a884ac 100644
--- a/.fernignore
+++ b/.fernignore
@@ -33,3 +33,52 @@ AGENTS.md
# ignore context7 config file
context7.json
+
+# SSE streaming fixes for POST /v1/audio/stream/with-timestamps. Each entry
+# below records what has to land upstream before it can be deleted again.
+
+# Passes `eventDiscriminator: "type"` to core.Stream, which routes the response
+# through iterSseEvents() instead of the defective iterDataMessages(). The live
+# API sends a populated `event: speech.chunk` name, so the discriminator is
+# correct. Without it, regen drops back to `{ type: "sse" }` and silently loses
+# multi-line `data:` concatenation, the trailing flush that carries speech.done
+# and its billable_characters_count, and CRLF handling.
+# Remove once SpeechifyInc/speechify-api declares the SSE event discriminator on
+# the with-timestamps endpoint (response-stream `format: sse` plus
+# `event-discriminator: type`) so the generator emits it.
+# COST: this freezes the whole audio resource client. Future API changes to any
+# audio endpoint will not land on regen while this entry is listed.
+src/api/resources/audio/client/Client.ts
+
+# Fixes abort handling in core.Stream. The generated version assigns a private
+# AbortController that nothing ever reads, and registers an abort listener that
+# closes over `this`, so the handler is a no-op and a reused AbortSignal retains
+# every Stream built from it. The shielded version holds the signal and polls
+# `aborted`, raising AbortError rather than ending iteration as a silent
+# truncated success.
+# This is generator template code, not API surface, so no change to
+# SpeechifyInc/speechify-api can fix it. Remove once the fix ships in
+# fernapi/fern-typescript-sdk (generated here at 3.70.1) and regen emits it.
+src/core/stream/Stream.ts
+
+# Covers the shielded Stream.ts. Holds an abort test that actually fails when
+# abort breaks (the generated one passed on its own `break` statement), plus
+# coverage for the SSE shape the with-timestamps endpoint really sends.
+# Remove together with src/core/stream/Stream.ts once the generator fix lands.
+tests/unit/stream/Stream.test.ts
+
+# Streaming example narrows on `item.type` and handles speech.error explicitly.
+# A mid-stream speech.error is yielded as a plain event, never thrown, because
+# the 200 is already committed by then. The generated example reported that as a
+# successful call with truncated audio. Also shows the
+# Buffer.from(audio, "base64") decode the SDK does not do for the caller.
+# Remove once the upstream endpoint example in SpeechifyInc/speechify-api
+# carries the same handling, or once the SDKs agree to raise speech.error from
+# the iterator, which is a cross-SDK decision affecting the Python SDK too.
+README.md
+
+# Same streaming example fix as README.md, in the generated endpoint reference.
+# COST: this freezes the entire reference. New endpoints, parameter changes and
+# description edits will not appear here on regen while this entry is listed.
+# Remove on the same upstream change as README.md.
+reference.md
diff --git a/README.md b/README.md
index 704bffd..ca8a6f4 100644
--- a/README.md
+++ b/README.md
@@ -122,9 +122,44 @@ const response = await client.audio.streamWithTimestamps({
voice_id: "geffen_32"
}
});
+
+const audioChunks: Buffer[] = [];
+
for await (const item of response) {
- console.log(item);
+ switch (item.type) {
+ case "speech.chunk":
+ // `audio` is Base64 and the SDK does not decode it for you. It is
+ // absent on a marks-only chunk, which the last chunk often is.
+ if (item.audio != null) {
+ audioChunks.push(Buffer.from(item.audio, "base64"));
+ }
+ // Mark times are absolute ms from the start of the synthesis, so
+ // they apply to the concatenated audio, not to this chunk.
+ for (const mark of item.speech_marks ?? []) {
+ console.log(mark.start_time, mark.value);
+ }
+ break;
+
+ case "speech.error":
+ // A failure after the stream has opened arrives as an event, NOT as
+ // a thrown error: the 200 status is already committed. The SDK does
+ // not raise it for you, so handle it explicitly — otherwise the loop
+ // ends normally and you treat truncated audio as a success.
+ throw new Error(`${item.error.code}: ${item.error.message}`);
+
+ case "speech.done":
+ // Terminal event. There is no `[DONE]` sentinel.
+ console.log("billable characters:", item.billable_characters_count);
+ console.log("audio duration (ms):", item.audio_duration_ms);
+ break;
+
+ default:
+ // Ignore unrecognized event types so new ones cannot break you.
+ break;
+ }
}
+
+const audio = Buffer.concat(audioChunks);
```
## File Uploads
diff --git a/reference.md b/reference.md
index c1966b8..7c3caf1 100644
--- a/reference.md
+++ b/reference.md
@@ -204,10 +204,45 @@ const response = await client.audio.streamWithTimestamps({
voice_id: "geffen_32"
}
});
+
+const audioChunks: Buffer[] = [];
+
for await (const item of response) {
- console.log(item);
+ switch (item.type) {
+ case "speech.chunk":
+ // `audio` is Base64 and the SDK does not decode it for you. It is
+ // absent on a marks-only chunk, which the last chunk often is.
+ if (item.audio != null) {
+ audioChunks.push(Buffer.from(item.audio, "base64"));
+ }
+ // Mark times are absolute ms from the start of the synthesis, so
+ // they apply to the concatenated audio, not to this chunk.
+ for (const mark of item.speech_marks ?? []) {
+ console.log(mark.start_time, mark.value);
+ }
+ break;
+
+ case "speech.error":
+ // A failure after the stream has opened arrives as an event, NOT as
+ // a thrown error: the 200 status is already committed. The SDK does
+ // not raise it for you, so handle it explicitly — otherwise the loop
+ // ends normally and you treat truncated audio as a success.
+ throw new Error(`${item.error.code}: ${item.error.message}`);
+
+ case "speech.done":
+ // Terminal event. There is no `[DONE]` sentinel.
+ console.log("billable characters:", item.billable_characters_count);
+ console.log("audio duration (ms):", item.audio_duration_ms);
+ break;
+
+ default:
+ // Ignore unrecognized event types so new ones cannot break you.
+ break;
+ }
}
+const audio = Buffer.concat(audioChunks);
+
```
diff --git a/src/api/resources/audio/client/Client.ts b/src/api/resources/audio/client/Client.ts
index c041972..c96bc8e 100644
--- a/src/api/resources/audio/client/Client.ts
+++ b/src/api/resources/audio/client/Client.ts
@@ -334,6 +334,7 @@ export class AudioClient {
signal: requestOptions?.abortSignal,
eventShape: {
type: "sse",
+ eventDiscriminator: "type",
},
}),
rawResponse: _response.rawResponse,
diff --git a/src/core/stream/Stream.ts b/src/core/stream/Stream.ts
index 8ccdecf..2ca31dc 100644
--- a/src/core/stream/Stream.ts
+++ b/src/core/stream/Stream.ts
@@ -46,7 +46,7 @@ export class Stream implements AsyncIterable {
private messageTerminator: string;
private streamTerminator: string | undefined;
private eventDiscriminator: string | undefined;
- private controller: AbortController = new AbortController();
+ private signal: AbortSignal | undefined;
private decoder: TextDecoder | undefined;
constructor({ stream, parse, eventShape, signal }: Stream.Args & { parse: (val: unknown) => Promise }) {
@@ -60,7 +60,10 @@ export class Stream implements AsyncIterable {
} else {
this.messageTerminator = eventShape.messageTerminator;
}
- signal?.addEventListener("abort", () => this.controller.abort());
+ // Held rather than subscribed to. An "abort" listener closes over `this`,
+ // so a long-lived reused signal would retain every Stream built from it;
+ // the read loop polls `aborted` instead and nothing outlives this object.
+ this.signal = signal;
// Initialize shared TextDecoder
if (typeof TextDecoder !== "undefined") {
@@ -179,8 +182,28 @@ export class Stream implements AsyncIterable {
return { [this.eventDiscriminator]: eventType, ...obj };
}
+ /**
+ * Halts iteration as soon as the caller's signal is aborted, so an aborted
+ * stream raises an AbortError instead of ending as a silent, truncated
+ * success. Rethrows the caller's own abort reason when one was given.
+ */
+ private throwIfAborted(): void {
+ if (this.signal?.aborted !== true) {
+ return;
+ }
+ const abortReason: unknown = this.signal.reason;
+ if (abortReason != null) {
+ throw abortReason;
+ }
+ const abortError = new Error("The stream was aborted");
+ abortError.name = "AbortError";
+ throw abortError;
+ }
+
async *[Symbol.asyncIterator](): AsyncIterator {
+ this.throwIfAborted();
for await (const message of this.iterMessages()) {
+ this.throwIfAborted();
yield message;
}
}
diff --git a/tests/unit/stream/Stream.test.ts b/tests/unit/stream/Stream.test.ts
index 83575f0..a6a6e6d 100644
--- a/tests/unit/stream/Stream.test.ts
+++ b/tests/unit/stream/Stream.test.ts
@@ -1,3 +1,4 @@
+import { vi } from "vitest";
import { Stream } from "../../../src/core/stream/Stream";
describe("Stream", () => {
@@ -366,6 +367,88 @@ describe("Stream", () => {
});
});
+ // The shape POST /v1/audio/stream/with-timestamps actually sends, verified
+ // against the live API: a populated `event:` name, a payload that already
+ // carries `type`, and a blank line after every frame.
+ describe("streamWithTimestamps SSE shape", () => {
+ const speechEventShape = { type: "sse", eventDiscriminator: "type" } as const;
+
+ async function collect(chunks: string[]): Promise {
+ const stream = new Stream({
+ stream: createReadableStream(chunks),
+ parse: async (val: unknown) => val,
+ eventShape: speechEventShape,
+ });
+
+ const messages: unknown[] = [];
+ for await (const message of stream) {
+ messages.push(message);
+ }
+ return messages;
+ }
+
+ it("should keep the payload's own type when the event name agrees", async () => {
+ const messages = await collect(['event: speech.chunk\ndata: {"type":"speech.chunk","audio":"QUJD"}\n\n']);
+
+ expect(messages).toEqual([{ type: "speech.chunk", audio: "QUJD" }]);
+ });
+
+ it("should keep the payload's own type when the event name is empty", async () => {
+ // The generated wire fixture still shows an empty `event:` name, so
+ // the discriminator must never overwrite an already-parsed type.
+ const messages = await collect(['event: \ndata: {"type":"speech.chunk","audio":"QUJD"}\n\n']);
+
+ expect(messages).toEqual([{ type: "speech.chunk", audio: "QUJD" }]);
+ });
+
+ it("should concatenate a data field split across multiple lines", async () => {
+ const messages = await collect([
+ 'event: speech.chunk\ndata: {"type":"speech.chunk",\ndata: "audio":"QUJD"}\n\n',
+ ]);
+
+ expect(messages).toEqual([{ type: "speech.chunk", audio: "QUJD" }]);
+ });
+
+ it("should yield a final event that has no trailing blank line", async () => {
+ // speech.done carries billable_characters_count. Dropping the last
+ // event because the stream ended without a terminator loses it.
+ const messages = await collect([
+ 'event: speech.done\ndata: {"type":"speech.done","billable_characters_count":31,"audio_duration_ms":1710}\n',
+ ]);
+
+ expect(messages).toEqual([{ type: "speech.done", billable_characters_count: 31, audio_duration_ms: 1710 }]);
+ });
+
+ it("should not treat a 'data:' substring inside a payload value as a field", async () => {
+ const messages = await collect([
+ 'event: speech.chunk\ndata: {"type":"speech.chunk","note":"data: not a field"}\n\n',
+ ]);
+
+ expect(messages).toEqual([{ type: "speech.chunk", note: "data: not a field" }]);
+ });
+
+ it("should ignore a non-data line containing a 'data:' substring", async () => {
+ const messages = await collect([
+ 'event: speech.chunk\n: keepalive data: ping\ndata: {"type":"speech.chunk","audio":"QUJD"}\n\n',
+ ]);
+
+ expect(messages).toEqual([{ type: "speech.chunk", audio: "QUJD" }]);
+ });
+
+ it("should read a chunk/done sequence across chunk boundaries", async () => {
+ const messages = await collect([
+ 'event: speech.chunk\ndata: {"type":"speech.chunk","audio":"QU',
+ 'JD"}\n\n\nevent: speech.done\ndata: {"type":"speech.done",',
+ '"billable_characters_count":31,"audio_duration_ms":1710}\n\n\n',
+ ]);
+
+ expect(messages).toEqual([
+ { type: "speech.chunk", audio: "QUJD" },
+ { type: "speech.done", billable_characters_count: 31, audio_duration_ms: 1710 },
+ ]);
+ });
+ });
+
describe("encoding and decoding", () => {
it("should decode UTF-8 text using TextDecoder", async () => {
const encoder = new TextEncoder();
@@ -447,7 +530,9 @@ describe("Stream", () => {
});
describe("abort signal", () => {
- it("should handle abort signal", async () => {
+ // No `break` in these loops on purpose: the only thing that can end
+ // iteration is the abort itself, so removing the abort fails the test.
+ it("should raise an AbortError and stop yielding when aborted mid-stream", async () => {
const controller = new AbortController();
const mockStream = createReadableStream(['{"value": 1}\n{"value": 2}\n{"value": 3}\n']);
const stream = new Stream({
@@ -458,17 +543,90 @@ describe("Stream", () => {
});
const messages: unknown[] = [];
- let count = 0;
- for await (const message of stream) {
- messages.push(message);
- count++;
- if (count === 2) {
- controller.abort();
- break;
+ let caught: unknown;
+ try {
+ for await (const message of stream) {
+ messages.push(message);
+ if (messages.length === 2) {
+ controller.abort();
+ }
+ }
+ } catch (error) {
+ caught = error;
+ }
+
+ expect(caught).toBeInstanceOf(Error);
+ expect((caught as Error).name).toBe("AbortError");
+ expect(messages).toEqual([{ value: 1 }, { value: 2 }]);
+ });
+
+ it("should yield nothing when the signal is already aborted", async () => {
+ const controller = new AbortController();
+ controller.abort();
+ const mockStream = createReadableStream(['{"value": 1}\n{"value": 2}\n']);
+ const stream = new Stream({
+ stream: mockStream,
+ parse: async (val: unknown) => val as { value: number },
+ eventShape: { type: "json", messageTerminator: "\n" },
+ signal: controller.signal,
+ });
+
+ const messages: unknown[] = [];
+ let caught: unknown;
+ try {
+ for await (const message of stream) {
+ messages.push(message);
+ }
+ } catch (error) {
+ caught = error;
+ }
+
+ expect((caught as Error).name).toBe("AbortError");
+ expect(messages).toEqual([]);
+ });
+
+ it("should propagate the caller's own abort reason", async () => {
+ const controller = new AbortController();
+ const reason = new Error("caller changed their mind");
+ const mockStream = createReadableStream(['{"value": 1}\n{"value": 2}\n']);
+ const stream = new Stream({
+ stream: mockStream,
+ parse: async (val: unknown) => val as { value: number },
+ eventShape: { type: "json", messageTerminator: "\n" },
+ signal: controller.signal,
+ });
+
+ const messages: unknown[] = [];
+ let caught: unknown;
+ try {
+ for await (const message of stream) {
+ messages.push(message);
+ controller.abort(reason);
}
+ } catch (error) {
+ caught = error;
}
- expect(messages.length).toBe(2);
+ expect(caught).toBe(reason);
+ expect(messages).toEqual([{ value: 1 }]);
+ });
+
+ it("should not subscribe to the signal, so a reused signal retains no streams", () => {
+ const controller = new AbortController();
+ const addEventListener = vi.spyOn(controller.signal, "addEventListener");
+
+ const streams = [1, 2, 3].map(
+ () =>
+ new Stream({
+ stream: createReadableStream([]),
+ parse: async (val: unknown) => val,
+ eventShape: { type: "json", messageTerminator: "\n" },
+ signal: controller.signal,
+ }),
+ );
+
+ expect(streams).toHaveLength(3);
+ expect(addEventListener).not.toHaveBeenCalled();
});
});
From 68982979ddf1d80952bc3647a02a27f3e17dd7db Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Wed, 19 Aug 2026 10:08:16 +0000
Subject: [PATCH 17/22] SDK regeneration
---
.fern/metadata.json | 8 ++++----
package.json | 2 +-
src/BaseClient.ts | 4 ++--
src/api/types/ErrorCode.ts | 4 ++++
src/version.ts | 2 +-
5 files changed, 12 insertions(+), 8 deletions(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 5821670..7eb75d1 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,9 +7,9 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "469722e153faa1a1c50aa63ea9cdca853f588721",
+ "originGitCommit": "cd9527ba61b3dd9af668b75ad5806f1774d6e7ee",
"originGitCommitIsDirty": false,
- "invokedBy": "manual",
- "requestedVersion": "4.0.0",
- "sdkVersion": "4.0.0"
+ "invokedBy": "ci",
+ "ciProvider": "github",
+ "sdkVersion": "3.0.2"
}
diff --git a/package.json b/package.json
index beff59d..0ab9286 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
{
"name": "@speechify/api",
- "version": "4.0.0",
+ "version": "3.0.2",
"private": false,
"repository": {
"type": "git",
diff --git a/src/BaseClient.ts b/src/BaseClient.ts
index 56fc834..2d206f1 100644
--- a/src/BaseClient.ts
+++ b/src/BaseClient.ts
@@ -63,8 +63,8 @@ export function normalizeClientOptions
Date: Wed, 19 Aug 2026 14:36:19 +0100
Subject: [PATCH 18/22] chore: restore 4.0.0 version pin after regeneration
rollback
The fern-api[bot] regeneration in commit 6898297 reverted the
version-bearing properties from 4.0.0 back to 3.0.2. This restores 4.0.0
in package.json, src/version.ts, .fern/metadata.json sdkVersion, and the
X-Fern-SDK-Version and User-Agent literals in src/BaseClient.ts.
This PR is a major release: it carries a "Release-As: 4.0.0" footer plus
three "BREAKING CHANGE:" footers in its body, which is what reaches main
on squash merge and drives the next version.
.release-please-manifest.json is deliberately left at 3.0.1. Setting it
to 4.0.0 would make release-please treat 4.0.0 as already released and
never cut the tag.
No files were added to .fernignore. package.json and .fern/metadata.json
must stay regenerable, and a future regen clobbering these strings again
is a known two-writer conflict between Fern and release-please, not
something to fix by shielding.
---
.fern/metadata.json | 2 +-
package.json | 2 +-
src/BaseClient.ts | 4 ++--
src/version.ts | 2 +-
4 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 7eb75d1..d17c0d4 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -11,5 +11,5 @@
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
- "sdkVersion": "3.0.2"
+ "sdkVersion": "4.0.0"
}
diff --git a/package.json b/package.json
index 0ab9286..beff59d 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
{
"name": "@speechify/api",
- "version": "3.0.2",
+ "version": "4.0.0",
"private": false,
"repository": {
"type": "git",
diff --git a/src/BaseClient.ts b/src/BaseClient.ts
index 2d206f1..56fc834 100644
--- a/src/BaseClient.ts
+++ b/src/BaseClient.ts
@@ -63,8 +63,8 @@ export function normalizeClientOptions
Date: Wed, 19 Aug 2026 14:41:55 +0100
Subject: [PATCH 19/22] Revert "chore: restore 4.0.0 version pin after
regeneration rollback"
This reverts commit d267778ba7d930aabc240ce04995cdd62ca0e53b.
The fern-api[bot] rollback to 3.0.2 in 6898297 was correct by design.
AGENTS.md is explicit that SDK_VERSION in src/version.ts and the
X-Fern-SDK-Version and User-Agent literals in src/BaseClient.ts are stamped
from the release tag in the CI checkout at publish time and never committed
back, so they are expected to read stale on the branch. Pinning them by hand
buys nothing and is called out as something not to do.
package.json version and .fern/metadata.json sdkVersion are owned by
release-please and set as part of the release PR. Both are addressed
structurally, by a known field and a jsonpath, so a regen cannot disarm them.
The Release-As: 4.0.0 footer in the PR body is what drives the next version.
---
.fern/metadata.json | 2 +-
package.json | 2 +-
src/BaseClient.ts | 4 ++--
src/version.ts | 2 +-
4 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index d17c0d4..7eb75d1 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -11,5 +11,5 @@
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
- "sdkVersion": "4.0.0"
+ "sdkVersion": "3.0.2"
}
diff --git a/package.json b/package.json
index beff59d..0ab9286 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
{
"name": "@speechify/api",
- "version": "4.0.0",
+ "version": "3.0.2",
"private": false,
"repository": {
"type": "git",
diff --git a/src/BaseClient.ts b/src/BaseClient.ts
index 56fc834..2d206f1 100644
--- a/src/BaseClient.ts
+++ b/src/BaseClient.ts
@@ -63,8 +63,8 @@ export function normalizeClientOptions
Date: Wed, 19 Aug 2026 13:44:31 +0000
Subject: [PATCH 20/22] SDK regeneration
---
.fern/metadata.json | 2 +-
src/api/types/ErrorCode.ts | 1 +
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 7eb75d1..4179b7e 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "cd9527ba61b3dd9af668b75ad5806f1774d6e7ee",
+ "originGitCommit": "5bad5fd56908ff01eecc7c80b8a1dc7a95ede569",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
diff --git a/src/api/types/ErrorCode.ts b/src/api/types/ErrorCode.ts
index 407d8f4..cc192a6 100644
--- a/src/api/types/ErrorCode.ts
+++ b/src/api/types/ErrorCode.ts
@@ -48,6 +48,7 @@ export const ErrorCode = {
AgentTestNotFound: "agent_test_not_found",
WorkspaceNotFound: "workspace_not_found",
InviteNotFound: "invite_not_found",
+ ProjectNotFound: "project_not_found",
InsufficientScope: "insufficient_scope",
PurchasedNumbersNotIncluded: "purchased_numbers_not_included",
PhoneNumberQuotaReached: "phone_number_quota_reached",
From 0a8cdacfa731a57caface14d53be6dbf0bdc3786 Mon Sep 17 00:00:00 2001
From: "fern-api[bot]" <115122769+fern-api[bot]@users.noreply.github.com>
Date: Wed, 19 Aug 2026 13:49:43 +0000
Subject: [PATCH 21/22] SDK regeneration
---
.fern/metadata.json | 2 +-
src/api/types/ErrorCode.ts | 1 +
2 files changed, 2 insertions(+), 1 deletion(-)
diff --git a/.fern/metadata.json b/.fern/metadata.json
index 4179b7e..79b62ef 100644
--- a/.fern/metadata.json
+++ b/.fern/metadata.json
@@ -7,7 +7,7 @@
"skipResponseValidation": true,
"namespaceExport": "Speechify"
},
- "originGitCommit": "5bad5fd56908ff01eecc7c80b8a1dc7a95ede569",
+ "originGitCommit": "d18d0fff37ec3446bf44071a930fcad840b961e6",
"originGitCommitIsDirty": false,
"invokedBy": "ci",
"ciProvider": "github",
diff --git a/src/api/types/ErrorCode.ts b/src/api/types/ErrorCode.ts
index cc192a6..1507fee 100644
--- a/src/api/types/ErrorCode.ts
+++ b/src/api/types/ErrorCode.ts
@@ -49,6 +49,7 @@ export const ErrorCode = {
WorkspaceNotFound: "workspace_not_found",
InviteNotFound: "invite_not_found",
ProjectNotFound: "project_not_found",
+ CrossProjectReference: "cross_project_reference",
InsufficientScope: "insufficient_scope",
PurchasedNumbersNotIncluded: "purchased_numbers_not_included",
PhoneNumberQuotaReached: "phone_number_quota_reached",
From dc7c92d509766fdfac27259cab143a604996ffd1 Mon Sep 17 00:00:00 2001
From: Luke Oliff <289678208+luke-speechify@users.noreply.github.com>
Date: Wed, 19 Aug 2026 15:20:32 +0100
Subject: [PATCH 22/22] fix: throw ContentTooLargeError on oversized audio
bodies
POST /v1/audio/speech and streamWithTimestamps return 413 payload_too_large for oversized bodies, but the audio client surfaced it as a generic SpeechifyError. Wire the typed ContentTooLargeError on the audio client to match the voices client. Client.ts is .fernignore-shielded so this survives regeneration.
---
src/api/resources/audio/client/Client.ts | 15 +++++++++++++++
1 file changed, 15 insertions(+)
diff --git a/src/api/resources/audio/client/Client.ts b/src/api/resources/audio/client/Client.ts
index c96bc8e..9f729bf 100644
--- a/src/api/resources/audio/client/Client.ts
+++ b/src/api/resources/audio/client/Client.ts
@@ -113,6 +113,11 @@ export class AudioClient {
throw new Speechify.ForbiddenError(_response.error.body as Speechify.Error_, _response.rawResponse);
case 404:
throw new Speechify.NotFoundError(_response.error.body as unknown, _response.rawResponse);
+ case 413:
+ throw new Speechify.ContentTooLargeError(
+ _response.error.body as Speechify.Error_,
+ _response.rawResponse,
+ );
case 429:
throw new Speechify.TooManyRequestsError(
_response.error.body as Speechify.Error_,
@@ -225,6 +230,11 @@ export class AudioClient {
throw new Speechify.ForbiddenError(_response.error.body as Speechify.Error_, _response.rawResponse);
case 404:
throw new Speechify.NotFoundError(_response.error.body as unknown, _response.rawResponse);
+ case 413:
+ throw new Speechify.ContentTooLargeError(
+ _response.error.body as Speechify.Error_,
+ _response.rawResponse,
+ );
case 429:
throw new Speechify.TooManyRequestsError(
_response.error.body as Speechify.Error_,
@@ -356,6 +366,11 @@ export class AudioClient {
throw new Speechify.ForbiddenError(_response.error.body as Speechify.Error_, _response.rawResponse);
case 404:
throw new Speechify.NotFoundError(_response.error.body as unknown, _response.rawResponse);
+ case 413:
+ throw new Speechify.ContentTooLargeError(
+ _response.error.body as Speechify.Error_,
+ _response.rawResponse,
+ );
case 429:
throw new Speechify.TooManyRequestsError(
_response.error.body as Speechify.Error_,