background-agents
An open-source background agents coding system
Background Agents (Open-Inspect) is an open-source autonomous coding agent system that operates in isolated sandboxes with full development tooling. Inspired by Ramp's internal Inspect platform, it enables teams to delegate coding tasks to AI agents that work in the background across multiple repositories, create pull requests with proper attribution, and integrate with Slack, GitHub, Linear, and webhooks. The system supports parallel child sessions, scheduled automations, multiplayer collaboration, and multiple AI providers including Anthropic Claude, OpenAI, xAI Grok, and OpenCode Zen.
Key Features
Use Cases
- 01Automating code reviews and fixes triggered by GitHub pull request mentions or Sentry error alerts
- 02Running scheduled maintenance tasks like dependency updates or code formatting across multiple repositories
- 03Delegating feature implementation from Linear issues with automatic PR creation and issue linking
- 04Collaborating on complex refactors with multiple team members in real-time multiplayer sessions
- 05Executing parallel testing or deployment validation across microservice repositories simultaneously
- 06Triggering automated investigation and triage workflows from external monitoring webhooks
Related Agents
View morehermes-agent
The agent that grows with you
agency-agents
A complete AI agency at your fingertips - From frontend wizards to Reddit community ninjas, from whimsy injectors to reality checkers. Each agent is a specialized expert with personality, processes, and proven deliverables.
openinterpreter
A coding agent for open models like Kimi K3
cline
Autonomous coding agent as an SDK, IDE extension, or CLI assistant.
background-agents — FAQ
What is Background Agents and what does it do?+
Background Agents (Open-Inspect) is an autonomous AI coding agent platform that runs development tasks in isolated sandboxes with full tooling. It connects to Slack, GitHub, Linear, and webhooks to accept coding tasks, executes them in the background using various AI models, and creates attributed pull requests.
How do I install and deploy Background Agents?+
Follow the setup guide in docs/SETUP_GUIDE.md and deployment instructions in docs/GETTING_STARTED.md. You'll deploy the control plane to Cloudflare Workers with Durable Objects, configure sandbox infrastructure (Modal, Daytona, E2B, or OpenComputer), and set up a GitHub App installation for repository access.
What AI models does Background Agents support?+
The system supports Anthropic Claude (Haiku, Sonnet, Opus, Fable), OpenAI GPT models via ChatGPT subscription OAuth, xAI Grok via SuperGrok subscription, and OpenCode Zen models (Kimi, MiniMax, Qwen, GLM). You can select models per-session with adjustable reasoning effort controls.
Is Background Agents designed for multi-tenant use?+
No, this is a single-tenant system designed for organizations where all users are trusted members with shared repository access. It uses a shared GitHub App installation without per-user repository validation, following Ramp's internal-tool architecture. Deploy behind SSO/VPN for security.
What API keys or subscriptions are required?+
You need a GitHub App installation for repository access and at least one AI provider: Anthropic API key, OpenAI access via existing ChatGPT subscription OAuth, xAI via SuperGrok subscription, or OpenCode Zen credentials. Optional integrations require Slack workspace tokens, Linear API keys, or Sentry webhook configuration.
Is Background Agents free and open source?+
Yes, the software is MIT-licensed and free to use. However, you'll incur costs for cloud infrastructure (Cloudflare Workers, sandbox backends like Modal/E2B/Daytona) and AI model API usage or subscriptions.
How do I install background-agents?+
Open the source repository on GitHub and follow its README. background-agents is a agent — MCP Agents Market links you directly to the official repo.
Is background-agents free?+
background-agents is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.