Repolode
Submit
trycompai

trycompai/crm

An open source CRM built around AI agents that research contacts, record evidence and run on their own schedule.

Open-source alternative to:SalesforceHubSpot

View on GitHubWebsite
Stars
11,488
Forks
1,681
Open issues
51
Last commit
last month

What is it?

Comp AI CRM is a customer relationship manager where an AI agent does the research work rather than sitting beside a form. The agent runs as its own deployment with a work queue and schedule, decides which contacts to look at next, books follow-ups, and stops when its research budget runs out. Each record is filled from observed evidence, and weak evidence is surfaced as a suggestion for a person to confirm rather than written as fact.

The agent works with no external API keys, reading the team's own email threads, meetings and signature blocks. Optional keys add web search and company and LinkedIn data. Reps can watch each contact's Agent tab to see the steps taken and talk to the agent in place. The app is a Next.js front end over a NestJS API with Postgres via Prisma, deployed on Vercel, with Google and Microsoft sign-in.

Key features

  • 01Autonomous research agent with its own work queue and schedule
  • 02Records written from observed evidence, with weak evidence held for human review
  • 03Works without external API keys by reading your own email, meetings and signature blocks
  • 04Agent tab on each contact, company and deal showing steps and questions
  • 05Google and Microsoft sign-in with an email allow-list
  • 06Type-safe tRPC link between the Next.js front end and NestJS API

Use cases

  • A small sales team that wants an agent to keep account and contact records current
  • Self-hosting a CRM with your own Postgres database and identity provider
  • Experimenting with agents that act on a schedule and write to a shared data store

Quick start

  1. 1

    Clone the repository

    git clone https://github.com/trycompai/crm.git && cd crm
  2. 2

    Copy the example env file and install dependencies

    cp .env.example .env && bun install
  3. 3

    Start Postgres, apply migrations and run the dev server

    docker compose up -d && bun run db:deploy && bun run dev

Things to know

  • Requires Bun and Docker, and setting up a Google or Microsoft OAuth app for sign-in.
  • Many agent features are optional but need extra API keys, such as Context for brand data and LinkedIn.
  • Single tenant only, with no organizations support.

Topics

#crm#ai-agent#typescript#nextjs#nestjs#postgres#prisma#trpc

FAQ

What is crm?
An open source CRM built around AI agents that research contacts, record evidence and run on their own schedule. Comp AI CRM is a customer relationship manager where an AI agent does the research work rather than sitting beside a form. The agent runs as its own deployment with a work queue and schedule, decides which contacts to look at next, books follow-ups, and stops when its research budget runs out. Each record is filled from observed evidence, and weak evidence is surfaced as a suggestion for a person to confirm rather than written as fact.
Is crm free to use? What license does it have?
crm is open source under the MIT license. Check the license terms before using it in commercial products.
Is crm still maintained?
Updated within 6 months. The last commit to crm was on Sep 11, 2026.
What can I use crm for?
A small sales team that wants an agent to keep account and contact records current Self-hosting a CRM with your own Postgres database and identity provider Experimenting with agents that act on a schedule and write to a shared data store
How do I install crm?
Clone the repository: git clone https://github.com/trycompai/crm.git && cd crm Copy the example env file and install dependencies: cp .env.example .env && bun install Start Postgres, apply migrations and run the dev server: docker compose up -d && bun run db:deploy && bun run dev
Where can I get crm?
The source code and installation instructions are on GitHub: https://github.com/trycompai/crm

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects