feat: one platform — email campaigns, templates, automations and flow…

[onlypromarketer]

…s inside the CRM

Combine wacrm (WhatsApp CRM) and listmonk (email engine) into a single product. listmonk runs headless behind the CRM's API and is never exposed publicly; every email capability is a native page or builder step in wacrm.

Email section

• Campaigns: list, compose, edit, preview, test-send, start/pause/cancel
• Lists: create, and sync CRM contacts in as subscribers (phone + contact id carried in attribs; unsubscribed lists never re-added)
• Subscribers: browse, search (SQL-injection-safe against listmonk's raw query parameter)
• Templates: CRUD + preview for transactional emails and newsletter layouts
• Settings: SMTP + sender identity, with the password-mask round-trip handled so a save never overwrites the stored secret with bullets

Email inside the WhatsApp builders

• Automations: new steps send_email and add_to_mailing_list
• Flows: new send_email node (canvas, edges, validation, engine)
• Both send via listmonk's tx API in external mode; contacts without an email are skipped, not failed
• Migration 040 widens the flow_nodes.node_type CHECK

Deployment (deploy/)

• docker-compose.yml: single hostname, Caddy auto-HTTPS, engine sealed off, recipient-facing unsubscribe/tracking routes proxied under the CRM domain
• docker-compose.local.yml: same stack on localhost
• setup-api-user.sh: creates listmonk's API user and performs the restart its in-memory user cache requires
• README: setup, local Mailpit testing, why listmonk is a separate service (AGPL vs MIT), why Mautic was evaluated and not added

Verified end to end: a signed inbound WhatsApp webhook started a flow and an automation that delivered two personalised emails through SMTP to a local inbox, added the contact to a mailing list, and skipped cleanly for an email-less contact. 876 tests pass.

Summary

What changed

Test plan

• npm run typecheck clean.
• npm run lint — no new errors beyond the pre-existing backlog.
• npm run build succeeds.
• Feature / fix manually exercised in the browser (or the reason it can't be).

Related

[ArnasDon]

Thanks for sharing this. Per CONTRIBUTING.md this repository is a template: forks are expected to diverge, and upstream only takes small, scoped fixes (security, correctness, docs) opened against current main. This PR is a whole-fork snapshot (rebrand/stack/deploy changes, colliding migration numbers, and it now conflicts with main), so I am closing it rather than reviewing file by file. If there is a specific bug fix in here that applies to the upstream template, please open it as its own small PR — those land quickly.

This one also replaces LICENSE and README with ProMarketer branding, which is fine for your fork but cannot be merged upstream.