What is Comp AI CRM?
Comp AI CRM is an open-source, agent-first CRM for customer research and record maintenance. Comp AI CRM treats the CRM as the place where an autonomous agent keeps its notes, rather than adding a chat interface to a conventional database. Sales representatives can inspect the agent’s work and resolve uncertain findings.
Comp AI CRM separates the research agent from the application. The agent runs on its own deployment, schedule and work queue, chooses what to investigate next and stops when its research budget runs out. Closing the browser does not stop that background work.
What Comp AI CRM can do
Comp AI CRM combines research tools with an evidence ledger that determines whether findings should update a record or become suggestions. Tools report observations instead of model-generated confidence scores. Strong evidence writes to the record, while weak evidence remains available for a human to settle.
- Comp AI CRM reads existing threads, meetings and signature blocks.
- Comp AI CRM researches people and enriches company records.
- Comp AI CRM schedules rechecks with reasons visible to representatives.
- Comp AI CRM displays agent steps and questions on customer records.
- Comp AI CRM supports optional web research and Context data.
- Comp AI CRM reviews discovered connection tools before agent access.
How Comp AI CRM works
Comp AI CRM uses eve for durable agent sessions, tools, skills, schedules and sandboxes. A dispatcher leases due work and starts sessions, while expired leases release work after failed runs. Agent sessions survive redeployment, and durable conversations remain available after a browser reload.
Comp AI CRM provides an Agent tab on contacts, companies and deals when both processes share the configured bridge secret. Workspace owners or administrators manage external services through Settings → Connections. Connections set maximum access levels, while deployed agents receive narrower grants identifying the exact tools they may use.
Who Comp AI CRM is for
Comp AI CRM serves sales representatives who need researched customer records and technical teams willing to operate an agent-first CRM. The repository describes a deliberately single-tenant application. Human review remains part of the workflow when evidence is weak or the agent cannot distinguish between possible identities.
- Comp AI CRM helps representatives inspect the evidence behind contact updates.
- Comp AI CRM supports customer research from existing correspondence.
- Comp AI CRM lets administrators restrict external agent actions.
- Comp AI CRM supports developers deploying a dedicated CRM installation.
Comp AI CRM access and local setup
Comp AI CRM is available through its public GitHub repository under the MIT license. The repository provides a local quick start using Bun and Docker, with environment configuration, database migrations and optional seed data. The default release branch contains the last tagged release; main contains unreleased work.
Comp AI CRM setup starts by cloning the repository, copying the example environment file and installing dependencies. Users start Postgres with Docker Compose, apply migrations and launch development services. Authentication requires Google or Microsoft configuration initially, with a custom identity provider available through Settings → SSO after access.
Comp AI CRM strengths and limitations
Comp AI CRM can operate without optional external research keys by examining the workspace’s existing history. The sandbox has neither network access nor database credentials, limiting what shell commands can reach. External service access also includes tool review and narrower per-agent permissions rather than relying solely on a workspace-wide connection.
- Comp AI CRM requires Bun, Docker and configuration for local setup.
- Comp AI CRM uses a deliberately single-tenant architecture.
- Comp AI CRM needs a bridge secret for Agent tab conversations.
- Comp AI CRM needs credentials to enable optional external research sources.
Comp AI CRM FAQ
What is Comp AI CRM?
Comp AI CRM is an open-source CRM designed for AI agents. Comp AI CRM runs a separate research agent that manages its own work queue, investigates customer information, schedules follow-ups and records findings according to the strength of the available evidence.
Is Comp AI CRM open source?
Comp AI CRM is an open-source project available on GitHub under the MIT license. Comp AI CRM includes local setup instructions using Bun and Docker, along with environment configuration, database migrations and optional seed data.
Does Comp AI CRM need external API keys?
Comp AI CRM can perform research without external API keys by reading existing CRM threads, meetings and signature blocks. Comp AI CRM supports optional additional sources, including web research and Context company and LinkedIn data, when the corresponding credentials are configured.
How does Comp AI CRM handle uncertain customer information?
Comp AI CRM uses an evidence ledger to distinguish record updates from suggestions. Strong evidence writes to the record, while weak evidence goes to human review. Comp AI CRM tools report observations rather than asking the model to assign its own confidence scores.
How does Comp AI CRM control agent integrations?
Comp AI CRM lets workspace owners and administrators manage supported services in Settings → Connections. Comp AI CRM supports read-only or read-and-write access, requires review of new or changed tools, and gives each deployed agent a narrower grant specifying exact tool names and fingerprints.
From the community
No reviews yet
Be the first to review this tool. What you write helps the next person choose.




