Assistable API Documentation
API reference for the Assistable BuildShip Migration backend
Assistable API
Self-hosted Fastify backend replacing BuildShip flows. All routes are fully migrated to Prisma/Neon — zero Firestore or Bubble dependency.
Base URL
https://api.assistable.aiFor local development:
http://localhost:3000Authentication
Most routes use Bearer token authentication via the INTERNAL_API_KEY environment variable:
Authorization: Bearer <INTERNAL_API_KEY>The message-event webhook route validates via location-level access tokens (SubAccount lookup by locationId).
Routes Overview
| Route | Method | Auth | Volume | Description |
|---|---|---|---|---|
/api/v1/message-event | POST | Location token | ~41M/week | GHL webhook — inbound/outbound message processing |
/api/v1/agent-run | POST | API Key | ~300K/week | AI agent response pipeline |
/api/v1/system-variables | POST | None | ~549K/week | Dynamic variable resolution for AI |
/api/v1/make-call | POST | None | ~661K/week | Outbound call initiation |
/api/v1/inbound-call | POST | None | ~3K/week | Inbound call routing |
/api/v1/voice-webhook | POST | None | ~6K/week | Telnyx voice call lifecycle events |
/api/v1/tool-proxy | POST | None | ~29K/week | ConvAI tool call proxy |
Architecture
All data flows through Prisma/Neon as the single source of truth. External services:
- GHL API — Contact/conversation management (external CRM)
- Telnyx — Voice calls, SMS
- OpenAI — Chat completions, embeddings
- Pinecone — Vector search (RAG)
- Ingestion API — Unified data pipeline (
api.assistable.ai)