Self-hostedCustomer Relationship Management (CRM)

Twenty

Self-hosted customer relationship management (crm) software.

Twenty is self-hostable software from the Awesome Selfhosted data index. Agents can inspect the source, deploy it in a controlled environment, and operate the browser UI; API or CLI support should be verified in project docs.

Agent access evidence

Browser

strong

Self-hosted web app can be operated through a browser once deployed.

Evidence

Account creation

partial

No vendor signup is required for self-hosted use, but deployment, admin setup, and local auth configuration are usually required.

Evidence

Pricing clarity

strong

Self-hosted/open-source licensing makes software cost inspectable; infrastructure cost depends on deployment.

Evidence

Docs quality

partial

Project website or source repository is listed for documentation and deployment review.

Evidence

Sandbox

strong

Self-hosted deployment supports local, staging, or disposable test environments before production use.

Evidence

Readiness score

Browser-assisted

Browser-assisted: 52/100. Strongest in agent safety; weakest in machine operability.

Machine operability4/25

Agent safety19.6/25

Agent readability15/20

Auth and setup10.5/15

Production reliability2.4/15

Best for

Self-hosted customer relationship management (crm) workflows, internal tools, and controlled deployment experiments.

Pricing

Self-hosted/open-source licensing listed as AGPL-3.0; infrastructure costs vary by deployment.

Auth

Auth and permissions depend on the deployed app configuration; platform metadata: Docker. Source code: https://github.com/twentyhq/twenty

Account creation

No vendor signup is required for self-hosted use, but deployment and admin setup are required.

Browser support

Browser UI can be operated after deployment; exact flows depend on the app configuration.

Limitations

Key limitations: generated from community metadata, not a full security review. Verify maintenance activity, license, deployment docs, auth model, backup and restore path, update process, API or CLI support, and data exposure before production agent automation. Browser automation can be brittle around UI changes, MFA, captchas, policy prompts, and hidden confirmation states. Store credentials outside source control and use least-privilege scopes, short-lived tokens, or restricted keys where the product supports them.