AnyCap

The execution layer for coding agents. Give Claude Code, Cursor, and Codex real-world capabilities — image, video, vision, music, web search, and publishing — inside the agents you already use. Install in plain English. No npm vs brew vs curl debate.

---

Quick start (30 seconds)

Paste this to Claude Code, Cursor, Codex, or any agent that can run a shell:

Read https://raw.githubusercontent.com/anycap-ai/anycap/main/llms.txt and follow the
instructions to install the AnyCap CLI and skill. If you can't access the URL, run:
  1. curl -fsSL https://anycap.ai/install.sh | sh
  2. npx -y skills add anycap-ai/anycap -s '*' -g -y
  3. anycap login
  4. anycap status

That's it. The agent installs the CLI, installs the skill, logs you in, and verifies the connection — then every capability below becomes a single sentence. Prefer to do it by hand? Jump to Manual install.

Contents

• What this fixes
• What your agent can do
• Install by talking to your agent
• Why skills over SDKs
• Manual install
• Capabilities
• Works with
• FAQ & troubleshooting
• Links
• License

---

What this fixes

Your coding agent can reason and write code. It still can't ship a meme, generate a hero image, read a screenshot a user dropped in, search the live web, or publish a landing page.

Without AnyCap	With AnyCap
"I can't generate images."	One command → a finished asset.
"I can't see this screenshot."	Reads images, video, and audio.
"My knowledge has a cutoff."	Live web search with citations.
"You'll need to deploy that yourself."	Ships a hosted page and returns the URL.

AnyCap is the missing execution layer — multimodal, web, and delivery behind one CLI and one auth.

What your agent can do

1. Make a meme end-to-end

Agent drafts the caption, generates the image, revises the visual, then returns a shareable link.

anycap image generate --prompt "cat in a tiny chef hat, meme caption 'ship it'" \
  --model nano-banana-2 -o meme.png
anycap image generate --prompt "make the caption bigger and yellow" \
  --model nano-banana-2 --mode image-to-image --param images=./meme.png -o meme-v2.png
anycap drive upload meme-v2.png --parent-path /memes
anycap drive share --src-path /memes/meme-v2.png

2. Turn a prompt into a hosted page

Agent writes the copy, generates hero visuals, and deploys a live URL.

anycap image generate --prompt "abstract product hero, soft gradient" \
  --model nano-banana-2 -o ./dist/hero.png
anycap page deploy ./dist --name "launch-page" --publish

3. Review a screenshot with a human in the loop

Agent opens the annotation UI, the human marks issues, the agent reads the feedback and revises.

anycap annotate ./screenshot.png --no-wait
anycap annotate poll --session ann_xxxx
anycap actions image-read --file ./screenshot-annotated.png \
  --instruction "List every change requested by the annotations"

Install by talking to your agent

AnyCap installs itself in plain English. No package-manager dance, no doc spelunking, no version pinning. Hand the Quick start paragraph to your agent — it reads llms.txt, installs the CLI, installs the skill, opens the browser to log you in, and verifies the connection.

After that one paste, everything else is just conversation. The skill teaches the agent the full command surface, so future actions become one sentence:

• "Upgrade AnyCap and check my status."
• "Generate a hero image for the launch page and deploy it."
• "Search the web for the latest Cursor changelog and summarize."

The agent picks the right capability, runs it, and hands you the result. You never touch a flag unless you want to.

Why skills over SDKs

Most tools ship an SDK and ask you to wire it in. AnyCap ships a skill — a single file your agent reads — plus a thin CLI it can call.

• No glue code. The agent learns the whole command surface from one SKILL.md; you don't import a library or manage API objects.
• Model- and agent-agnostic. The same skill works across Claude Code, Cursor, Codex, and 40+ agents. Swap agents, keep your capabilities.
• One auth, one CLI. Image, video, vision, music, web, and publishing sit behind a single login instead of a dozen API keys.
• Plain-English invocation. Capabilities are triggered by intent, not by memorizing flags.

Manual install

1. Install the CLI

macOS / Linux / Windows (Git Bash):

curl -fsSL https://anycap.ai/install.sh | sh

npm (all platforms):

npm install -g @anycap/cli

Or grab a binary from GitHub Releases.

2. Install the skill

Works with Claude Code, Cursor, Windsurf, OpenCode, and 40+ agents:

npx -y skills add anycap-ai/anycap -s '*' -g -y

More ways to install the skill

# Direct download
curl -fsSL https://raw.githubusercontent.com/anycap-ai/anycap/main/skills/anycap-cli/SKILL.md \
  --create-dirs -o ~/.agents/skills/anycap-cli/SKILL.md

# Via AnyCap CLI
anycap skill install --target ~/.agents/skills/anycap-cli/

# Check if the skill is up to date
anycap skill check --target ~/.agents/skills/anycap-cli/SKILL.md

3. Verify

anycap login
anycap status

Capabilities

Capability	Command	What agents do with it
Image generation / edit	anycap image generate (--mode image-to-image for edits)	Hero art, meme assets, illustration, photo edit
Image / video / audio reading	anycap actions image-read / video-read / audio-read	Screenshot review, meeting transcripts, QA
Video generation	anycap video generate	Demo clips, animated assets
Music generation	anycap music generate	Jingles, soundtracks
Web search	anycap search --query (general) / --prompt (grounded with cites)	Live answers, research
Web crawl	anycap crawl <url>	Any URL into clean Markdown
Annotate	anycap annotate	Human-in-the-loop visual feedback
Draw	anycap draw	Live whiteboard / Mermaid diagrams
Drive	anycap drive upload / share	Shareable file links
Page	anycap page deploy	Hosted static pages
Download	anycap download	Save any remote file

Coming soon: TTS / voice synthesis.

Models

AnyCap's model lineup evolves quickly, so the authoritative, always-current list lives in the CLI — run anycap <capability> models to see exactly what your account can use:

anycap image models   # image generation / edit
anycap video models   # video generation
anycap music models   # music generation

Inspect a single model's modes and parameter schema before calling it:

anycap image models nano-banana-2                                # modes + operations
anycap image models nano-banana-2 schema --mode text-to-image    # parameter schema

The tables below reflect the models available in CLI v0.3.6 — newer versions add and retire models, so treat the CLI output as the source of truth.

Image generation / edit

Model	Name	Modes
nano-banana-2	Nano Banana 2	text-to-image, image-to-image
nano-banana-pro	Nano Banana Pro	text-to-image, image-to-image
gpt-image-2	GPT Image 2	text-to-image, image-to-image
flux-kontext-max	FLUX.1 Kontext Max	text-to-image, image-to-image
qwen-image	Qwen Image	text-to-image, image-to-image
seedream-4.5	Seedream 4.5	text-to-image, image-to-image
seedream-5	Seedream 5	text-to-image, image-to-image

Video generation

Model	Name	Modes
veo-3.1	Veo 3.1	text-to-video, image-to-video
veo-3.1-fast	Veo 3.1 Fast	text-to-video, image-to-video
sora-2-pro	Sora 2 Pro	text-to-video, image-to-video
kling-3.0	Kling 3.0	text-to-video, image-to-video
kling-o1	Kling O1	image-to-video
hailuo-2.3	Hailuo 2.3	text-to-video, image-to-video
seedance-1.5-pro	Seedance 1.5 Pro	text-to-video, image-to-video
seedance-2	Seedance 2.0	text-to-video, image-to-video
seedance-2-fast	Seedance 2.0 Fast	text-to-video, image-to-video

Music generation

Model	Name	Modes
suno-v5	Suno V5	text-to-music
suno-v5.5	Suno V5.5	text-to-music
mureka-v8	Mureka V8	text-to-music
elevanlabs-music	ElevenLabs Music	text-to-music

Reading / understanding (anycap actions image-read / video-read / audio-read) automatically selects the best available model. Pass --model only when you need a specific one.

Works with

Claude Code · Cursor · Codex · Windsurf · OpenCode · 40+ agents via skills.sh.

FAQ & troubleshooting

Do I need an API key for each capability?

No. One anycap login covers every capability — image, video, vision, music, web, and publishing share a single auth.

The agent says it can't reach llms.txt.

Run the four manual commands from Quick start, or follow Manual install.

The skill isn't showing up in my agent.

Confirm it installed into your agent's skills directory, then re-check:

anycap skill check --target ~/.agents/skills/anycap-cli/SKILL.md

Reload your agent so it re-reads the skills directory.

How do I update?

Tell your agent "Upgrade AnyCap and check my status," or re-run the install script and npx -y skills add anycap-ai/anycap -s '*' -g -y.

Which platforms are supported?

macOS, Linux, and Windows (via Git Bash). Install through the script, npm, or a binary from Releases.

Links

• llms.txt — Give this to your agent
• Skill file — Full capability documentation
• GitHub Releases — CLI binaries
• skills.sh — Skills directory listing
• Website

License

MIT