manageCallAI
Health Uyari
- License — License: Apache-2.0
- Description — Repository has a description
- Active repo — Last push 0 days ago
- Low visibility — Only 6 GitHub stars
Code Gecti
- Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Gecti
- Permissions — No dangerous permissions requested
Bu listing icin henuz AI raporu yok.
AI-native telecom control plane over FreeSWITCH — IVR, routing, PBX, bounded AI operator workflows (v0.6)
manageCallAI
AI-native telecom control plane over FreeSWITCH with REST, MCP, n8n, visual IVR, validation, simulation, rollback, runtime safety controls, and bounded AI operator workflows.
manageCallAI is an open-source platform for building programmable PBX, IVR, and telecom automation systems on top of stock FreeSWITCH. The control plane lives in TypeScript and PostgreSQL. FreeSWITCH remains runtime-only. Lua remains thin. The Go agent handles ESL/runtime integration. v0.6 introduces assistive AI workflows grounded in API-owned records — call failure explanation, route risk analysis, recording summary review, and natural-language reporting — without autonomous runtime control.
Canonical docs
- docs/architecture/source-of-truth.md
- docs/architecture/overview.md
- docs/architecture/runtime-boundaries.md
- docs/design/software-design.md
- docs/design/domain-model.md
- docs/design/database-schema.md
- docs/design/setup-bootstrap.md
- docs/README.md
Implemented in the repository
Source inspection shows implemented support for:
- multi-tenant auth and tenant-scoped PBX objects
- extensions, trunks, phone numbers, schedules, inbound/outbound routes
- queues, call groups, voicemail boxes, prompt assets, recordings, and call events
- IVR draft, validate, simulate, publish, rollback, and approval-aware flows
- FreeSWITCH directory and dialplan callbacks over
mod_xml_curl - Go FreeSWITCH agent runtime/event integration
- tenant and platform runtime visibility surfaces
- feature codes, parking, conference rooms, runtime apply requests, and end-user self-service
- first-run setup/bootstrap through
/setupor headlessSETUP_*environment variables - deployment packaging with
docker-compose.prod.yml,install.sh, and a Helm chart scaffold - v0.6 AI operator workflows: bounded call failure explanation, route and change risk analysis, recording/voicemail summary review, and natural-language telecom reporting
Release posture
Current release: v0.6.2 (2026-06-06) — combined AI-expansion + end-user comms release. v0.6.1 scope merged into v0.6.2.
Evidence: docs/release/release-evidence-v0.6.2.json
Release posture must be derived from release evidence, not from source inspection alone.
- Implementation in the tree is not evidence.
- Scripts, templates, and
--check-configmodes are not evidence. - Production claims require artifacts tied to the release-candidate commit and workflow runs.
Use these documents for current stage and blockers:
- docs/release/release-checklist.md
- docs/release/product-release-audit.md
- docs/ops/release-process.md
- docs/planning/open-release-blockers.md
Architecture summary
React Web UI
-> REST API
-> PostgreSQL desired state
-> validation / simulation / publish / rollback / audit
-> runtime artifact generation
-> FreeSWITCH mod_xml_curl directory/dialplan
-> Lua thin executor
-> Go ESL agent
-> call events / observability
MCP / n8n
-> safe API abstractions only
Quick start
pnpm install
pnpm db:up
pnpm db:migrate
pnpm --filter @managecallai/api dev
For packaging and first boot:
manageCallAI is not production-ready unless all production evidence gates pass with real artifacts tied to the release candidate commit.
Yorumlar (0)
Yorum birakmak icin giris yap.
Yorum birakSonuc bulunamadi