Get This Tool
Atuin AI Proxy
Pricing
- Model
- Free
Summary
Atuin's AI features are enabled by default expecting a specific Hub endpoint — if you're not routing through Atuin's own cloud, those requests go nowhere and the feature silently breaks. atuin-ai-proxy exists to solve that exact gap.
The proxy intercepts the endpoint Atuin expects, then forwards requests to any OpenAI-compatible backend you configure — your own Ollama instance, a self-hosted model, GitHub Codex, or the upstream OpenAI API. You point Atuin's config at localhost, set a token, and the proxy handles the translation. Docker Compose and a local Python runner are both supported. Debugging is addressed through request ID logging, so when a shell completion returns garbage you have a thread to pull. The tool is a thin forwarding layer — it does not add caching, rate limiting, or response modification.
Bottom line: Pick this when you need Atuin's AI shell features running against your own backend without touching Atuin's cloud; skip it when you need anything beyond raw request forwarding, like response filtering or cost controls, because none of that is here.
Community Performance Report Card
No community ratings yet. Be the first to rate this tool!
Community Benchmarks Community
Sign in to submit a benchmarkNo community benchmarks yet. Be the first to share a real-world data point.
Pros
Sign in to edit- Backend selection via a single environment variable, so switching from OpenAI to a self-hosted model when API costs or privacy requirements change is a config file edit — not a redeployment or code change.
- Ships with Docker Compose support out of the box, which means you avoid writing your own container setup to get the proxy running in a reproducible environment.
- Request ID logging is included, so when Atuin's AI completions return unexpected results you have a correlation handle to trace the failure back through your backend's logs.
- Token-based authentication is configurable but optional, which means solo local setups skip the overhead while networked deployments can add a bearer token without modifying the proxy code.
- Open-source with no paid tiers or vendor lock-in, so the full forwarding logic is auditable — relevant when you are routing shell history context through any intermediary.
Cons
Sign in to edit- The proxy does no response modification, caching, or rate limiting — teams that need cost controls or abuse prevention in front of their model backend have to build that layer themselves, and at that point they are running two systems.
- There is no multi-tenant token management: one proxy instance, one backend, one token. Teams supporting multiple users or multiple Atuin installations against different backends will need separate proxy instances per configuration, which does not scale.
- The project has 3 stars and 0 open issues as of the scraped snapshot — not because it is perfect, but because the user base is small. When something breaks at the integration layer between an Atuin update and the proxy's endpoint assumptions, community debugging resources are precious little, and teams with a support SLA requirement will switch to a maintained API gateway with an Atuin-compatible route instead.
Community Reviews
Sign in to write a reviewNo reviews yet. Be the first to share your experience.
About
- API Available
- Yes
- Self-Hosted
- Yes
- Last Updated
- 2026-07-12T08:15:55.246Z
Best For
Who it's for
- Atuin users wanting AI shell assistance
- Developers needing flexible backend routing for Atuin
- Self-hosted environments avoiding direct cloud API exposure
What it does well
- Enable Atuin AI features locally
- Route Atuin requests through custom OpenAI-compatible APIs
- Debug Atuin AI interactions with request IDs and detailed logging
Integrations
Discussion Community
Sign in to commentNo discussion yet. Sign in to start the conversation.
Compare Atuin AI Proxy
Spotted incorrect or missing data? Join our community of contributors.
Sign Up to ContributeCommunity Notes & Tips Community
Sign in to contributeBe the first to contribute. General notes, observations, gotchas, and tips from people who use this tool day-to-day.
Frequently Asked Questions
- Is Atuin AI Proxy free?
- Yes — Atuin AI Proxy is fully free to use. There is no paid tier.
- Is Atuin AI Proxy open source?
- Yes. Atuin AI Proxy is open source.
- Does Atuin AI Proxy have an API?
- Yes. Atuin AI Proxy exposes a developer API. See the official documentation at https://github.com/vlad1mir-d/atuin-ai-proxy for details.
- Can I self-host Atuin AI Proxy?
- Yes. Atuin AI Proxy supports self-hosting on your own infrastructure.
Hours Saved & ROI Stories Community
Sign in to contributeBe the first to contribute. Concrete time/cost savings, with context. e.g. "Cut my code review backlog from 4h to 45m per week."
Best Atuin AI Proxy alternatives →
Curated lists that include this category
atuin-ai-proxy is a Python 3 proxy that exposes the specific Hub AI endpoint Atuin’s client expects, then forwards those requests to an OpenAI-compatible Chat Completions or Responses backend of your choosing. The core workflow is three steps: copy the example environment file, set your backend credentials and model, and point Atuin’s config at the proxy’s local address with a matching token. From there, Atuin behaves as if it is talking to its native cloud endpoint.
The proxy supports three distinct backend modes, which is where it earns its flexibility over a plain reverse proxy. You can route to any OpenAI-compatible API with a standard key, authenticate to GitHub Codex via an access token and account ID, or use a Codex OAuth auth.json file — each mode is selected by a single environment variable. That means swapping from a hosted model to a self-hosted one is an environment file edit, not a code change.
This tool fits one scenario well: a developer running Atuin in a self-hosted or air-gapped environment who wants AI shell assistance without sending requests to Atuin’s cloud. It breaks — or rather, stops being sufficient — the moment you need anything the proxy does not do: response caching, per-user rate limiting, content filtering, or multi-tenant token management. Teams with those requirements end up wrapping the proxy in additional infrastructure, at which point the question of whether to use a more capable API gateway becomes real.
The repo ships a Dockerfile, a docker-compose example, and a pyproject.toml, so deployment paths are covered for both containerized and direct Python environments. If ATUIN_PROXY_TOKEN is left unset, the proxy accepts local requests without bearer authentication — convenient for solo developer setups, a deliberate risk in any networked environment.