Verified project record

trycompai/crm

crm is an open-source, agent-first CRM where a durable research agent maintains CRM records, performs scheduled research and follow-ups, and records evidence-backed facts with human review of weak evidence. It ships a Next.js front end, NestJS API, authentication, mailbox synchronization, and a Postgres-backed CRM application.

Project overview

Instead of bolting a chat feature onto a conventional form-based CRM, this project makes the agent the center of the CRM, distinguishing strong evidence written to records from weak evidence routed to human review rather than treating model confidence as truth.

Project type: AI Agent · Workflow
Use cases: Automation
Deployment: Refer to project documentation
License: MIT

Best for

Developers and teams who want an open-source, agent-first CRM where the agent itself researches, verifies, and maintains customer records.
Teams that want agent-written facts to be evidence-priced, with weak evidence routed to human review instead of automatic assertion.

Key capabilities

Records observed facts in CRM records, prices the evidence, and turns weak evidence into suggestions for human settlement rather than treating model confidence as truth.
Ships a Next.js front end, NestJS API, authentication, mailbox synchronization, and Postgres-backed CRM application.
Lets the agent search the open web with citations when a Perplexity API key is configured.

Limitations and risks

The codebase deliberately supports a single tenant and has no organizations.
External source integrations are optional, so functionality from those sources is unavailable when their keys are absent.
Deployment consists of three independent services plus Postgres: the Next.js app, NestJS API, and agent.
The application uses Vercel AI Gateway for its model layer, so running it depends on this external service; a local-only data boundary is not established.
At least one identity provider is needed for sign-in, with Google, Microsoft, or a user-provided identity provider documented.

Getting started

The documented local setup requires Bun, Docker, environment configuration, dependency installation, database startup, migrations, and application startup; authentication may require OAuth credentials or another identity provider. Steps: clone the repository and change into it; copy .env.example to .env and fill in values; run bun install; run docker compose up -d; run bun run db:deploy; run bun run dev; open http://localhost:3000.

Evidence and sources

GitHub project description: Comp AI CRM is an open source, CRM designed for AI agents. Agentic-first CRM.
README: Most CRMs are a database with a form in front of it. The AI ones bolt a chat box onto the side of that form. Both leave the actual work — finding out what is true, and writing it…
README: The rule the agent itself never breaks: **nothing about a person is guessed.** No tool accepts a confidence score, because a model asked to grade its own certainty will, and it wi…
README: It decides what to look at next, books its own follow-ups, spends a research budget, and stops when the budget runs out. Nothing about it is request-response: close the browser an…
README: You need [Bun](https://bun.com) and Docker.

Source updated at: 2026-09-22T18:07:39Z

Open original GitHub project