11ai Vercel Chat SDK v4 integrations

Name both systems, data and actions crossing the seam, trust boundary, persistence owner, credential source, and production target before editing.

Version baseline: Target Chat SDK 4, currently the 4.35 release line. Use chat/ai for AI utilities, current adapters, Vercel Connect helpers, distributed state, webhook deduplication, and current platform capability checks.

Inspect the seams

node -p "require('chat/package.json').version" 2>/dev/null
rg -n 'new Chat|create.*Adapter|onNewMention|onSubscribedMessage|chat/ai|@vercel/connect/chat|webhooks\.|create.*State' . --glob '*.{ts,tsx,js}' --glob '!node_modules' | head -140
rg -o '^[A-Z][A-Z0-9_]*=' .env.example .env.local 2>/dev/null | sed 's/=.*//' | sort -u
rg -n "webhook|route|database|redis|telemetry|deploy|workflow|sandbox" . --glob '!node_modules' | head -100

Find existing adapters and middleware before adding another. Keep this plugin standalone and document external contracts locally.

Wire deliberately

Read references/integrations.md for standalone runtime, persistence, auth, observability, testing, and deployment patterns.

Change one seam. Keep bot tokens, signing secrets, app passwords, webhook secrets, connector credentials, message content, or personal data server-side, validate untrusted inputs, minimize retained content, and make externally visible actions explicit.

Verify end to end

npm run typecheck --if-present
npm test --if-present
npm run build --if-present

Test a bounded success, rejection, timeout, duplicate, and recovery path. Use test targets or mocks unless the user approves a real external call.

Report

State systems, trust and persistence boundaries, credentials by name, externally visible actions, data retention, budget or rate limits, files, checks, and rollback. Hand failures to 11ai-operator-vercel-chat-sdk-v4-troubleshooting.