Deputies
an open source background agent
Last verified:
What is Deputies?
Deputies is a Flue-powered background-agent control plane designed for engineering teams who want to delegate real coding work to AI agents while maintaining full visibility and control. It runs on your existing infrastructure with a Node API, Postgres database, and static React frontend, allowing you to run agents in sandboxes tailored to your development environment with the right repositories, tools, services, and context sources attached.
The tool enables you to assign work by sending prompts from the web UI, Slack, GitHub, or webhooks. Deputies then manages the sandbox, work queue, logs, services, and artifacts while the agent runs asynchronously. You can review outputs by inspecting transcripts, opening tools, downloading artifacts, and following up. Finally, you can ship or iterate by queuing follow-ups, publishing pull requests, or merging finished work.
Deputies is built for engineering teams and developers who want to hand off authentic engineering tasks to AI coding agents without losing visibility, access control, history, callbacks, and deployment concerns in one place. It supports multiple model providers including OpenAI Codex with ChatGPT Pro subscription, runs sandboxes on Daytona or Docker, and can be deployed anywhere Node and Postgres can run, with Railway and Docker Compose docs available.
Key features include starting background engineering sessions from multiple interfaces, assigning follow-ups and queuing additional work, watching the agent's work trail in real-time, persisting and reviewing diagnostics and generated files, opening app previews from sandbox services, inspecting or editing output in a browser-based IDE and diff viewer, and creating PRs from completed work.
Deputies pricing
Pricing model: Freemium
The website does not explicitly mention pricing details. Deputies is open source and self-hostable, running on your existing infrastructure. It works with ChatGPT Pro through @earendil-works/pi-ai/oauth since Flue uses Pi under the hood. You need your own LLM subscription for model access, and you must provide your own infrastructure (Node, Postgres, sandbox provider like Daytona or Docker).
Deputies pros
- Runs on your existing infrastructure with Node API and Postgres
- Open source with self-hostable deployment
- Supports multiple model providers including OpenAI Codex
- Start sessions from web UI, Slack, GitHub, or webhooks
- Browser-based IDE and diff viewer for editing output
- Run sandboxes on Daytona or Docker
- Full visibility into agent transcripts and logs
- Create pull requests directly from completed work
- Persist and review diagnostics and generated artifacts
- Open app previews from sandbox services
- Assign follow-ups and queue additional work easily
- Deploy anywhere Node and Postgres can run
- Railway and Docker Compose deployment docs available
- Explicit boundaries for web UI, API, worker, database
- Can scale from monolith to multi-service deployment
Deputies cons
- Recurring and proactive tasks not yet available
- RBAC for teams not implemented yet
- Traces for deeper run inspection coming soon
- Evals for measuring agent quality not available
- Tagging and filtering needs improvement
- Session management features are limited
- Richer reporting and review workflows pending
- Snapshot and restore workflows not ready
Frequently asked questions about Deputies
What is a background agent?
A background agent is an AI coding agent that runs asynchronously instead of sitting inside your terminal. You hand it a task, it works in its own environment, and you come back to review progress, artifacts, logs, callbacks, and the final result. Deputies is built for teams that want to hand off real engineering work to agents while keeping visibility, access control, history, callbacks, and deployment concerns in one place.
What other systems like this exist?
The space includes open-source projects like Open-Inspect and Open SWE, hosted or enterprise-oriented products like OpenHands, Ona, and Devin, and vendor-native cloud agents like Claude Code and OpenAI Codex. Deputies is another take on the same broad category, with a stronger focus on self-hostable deployment alongside your existing infrastructure.
Why would I want my background agent to be open source?
Background agents need access to your code, secrets, development environments, issue trackers, and deploy workflows. Open source lets you self-host the control plane inside the same security boundary as your other systems, while keeping it inspectable, auditable, and adaptable to the deployment patterns your team already uses. It also gives you an escape hatch to change integrations, add providers, debug the full system, and connect to non-standard internal systems with custom adapters.
Does it work with my LLM subscription?
Deputies works with ChatGPT Pro through @earendil-works/pi-ai/oauth, because Flue uses Pi under the hood. Any LLM provider supported by Pi should be supported by Deputies, or at least be easy to add.
Why another background agent?
After working with and extending Open-Inspect and Open SWE, the creator was excited by the background-agent workflow but frustrated by their deployment constraints. Open-Inspect is tied to Cloudflare, while Open SWE pushes users toward LangChain's hosted LangSmith platform. When Flue came out, it looked like the right foundation for a background agent built with deployability front and center, running wherever the rest of your infrastructure already runs.
What interfaces can I use to start a session?
You can start a background engineering session from the web UI, Slack, GitHub, or custom webhooks. This gives you flexibility to assign work from wherever your team already communicates and manages projects.
What sandbox providers does Deputies support?
Deputies supports running sandboxes on Daytona or Docker. This gives you options depending on your existing infrastructure and sandboxing preferences.
Can I deploy Deputies on cloud platforms?
Yes, Deputies can be deployed anywhere Node and Postgres can run. Railway and Docker Compose docs are available to help with deployment, making it flexible for various cloud and on-premise environments.
What can I do with completed agent work?
You can queue a follow-up task, publish a pull request directly from the completed work, or merge the finished work. You can also inspect the transcript, download artifacts, review diagnostics and logs, and open app previews from sandbox services.
What features are coming soon to Deputies?
Coming soon features include recurring and proactive tasks, reusable agent prompts for common workflows, snapshot and restore workflows for moving work between remote and local sandboxes, RBAC for teams/repositories/agent access, traces for deeper run inspection, evals for measuring agent quality over time, improved tagging/filtering/session management, and richer reporting and review workflows.