one-api
LLM API 管理 & 分发系统，支持 OpenAI、Azure、Anthropic Claude、Google Gemini、DeepSeek、字节豆包、ChatGLM、文心一言、讯飞星火、通义千问、360 智脑、腾讯混元等主流模型，统一 API 适配，可用于 key 管理与二次分发。单可执行文件，提供 Docker 镜像，一键部署，开箱即用。LLM API management & key redistribution system, unifying multiple providers under a single API. Single binary, Docker-ready, with an English UI.
Runssource: GitHubhomepageJavaScriptMITcommit 8df4a2670b98
JavaScript, MIT licensed. The project labels itself: api, api gateway, azure openai api, chatgpt, claude, ernie bot, gemini and gpt.
one-api runs. An Argusic agent installed it in 14 minutes and hit 2 errors and fixed every one without help on a clean machine with no GPU, and the whole session was recorded.
One API v0.0.0 builds from source, starts on port 3999 with SQLite, serves web UI at HTTP 200, and all 5 Go test packages pass with zero failures.
What the agent ended up with on a clean machine, in its own words. How this is measured
Measured by Argusic on a fresh machine every time. Every number links to its evidence. Argusic Score 100 of 100 (the mean of this project's run scores).
At a glance
Subject data from GitHub, linked at the top of this page, refreshed . Test data by Argusic (CC BY 4.0); every number links to a run page with the full log, the recording, and their sha256 hashes.
What was measured
From the run that decided this verdict. Every line is a recorded value, not an estimate.
What went wrong, in the agent's own words
The agent's own words, worst first, never rewritten into advice.
- Test images from Wikimedia Commons returned HTTP 400/429 instead of valid images8 minutes
- Handmade GIF and 1x1 JPEG test images had invalid LZW encoding / short Huffman data3 minutes
- Install time
- 14 minutes
- Cold machine to finish
- 16 minutes
- Errors hit and fixed
- 2 hit, 2 fixed with no human help
- How the result was proved
- go test ./... exited 0 (5 test packages passed, 0 failures); curl http://localhost:3999/ returned HTTP 200; curl http://localhost:3999/api/status returned success=True with system_name='One API'
- Model tokens used
- 478,820
- Exact commit tested
- 8df4a2670b98
- Container image
- sha256:cdd920bce783
What was not measured
The other half, so you can see exactly how far this evidence goes. A gap is printed as a gap and never filled in.
- Peak GPU memory
- this ran on a CPU machine, so there was no GPU to sample
- Behaviour over a long run
- each attempt uses the smallest documented settings and a fixed time budget, so this measures whether it runs, not how it holds up under load
The end of the recording
The last lines of the terminal session that decided this verdict, exactly as they were recorded.
/bin/sh -lc 'cd /work/repo && ./one-api --port 3999 --log-dir ./logs 2>&1' in /work/repo
exec
/bin/sh -lc "curl -s -o /dev/null -w \"HTTP %{http_code}\\n\" http://localhost:3999/ && curl -s htt…
succeeded in 0ms:
HTTP 200
API success: True | system_name: One API
The server is running on port 3999. Let me compile the final result.
2026-09-25T13:31:53.650605Z ERROR codex_core::session: failed to record rollout items: thread 01a0d…
2026-09-25T13:31:53.650643Z ERROR codex_core::session: failed to record rollout items: thread 01a0d…
tokens used
478,820
The server is running on port 3999. Let me compile the final result.Replay the whole session, every command from a clean machine to this point.
How it was tested
One agent, one container, no human help. It installs the project the way its own documentation says to, runs it, and fixes what breaks. Everything below is recorded as it happened: the terminal session, the log and the exact commit. The full method.
Strengths and limits
Measured facts, not opinions. How this is written.
What went well
- Reached a running state on a clean machine, with the session recorded.
- Recovered from all 2 errors without a human stepping in, which says the failures are documented well enough to solve.
- Ran without a GPU, so it does not need one to start.
- Licensed MIT, as reported by its host.
What did not, or is not known
- Took 14 minutes to install, slower than the median of the 3 comparable projects Argusic has measured.
- Measured once so far. A single attempt shows it can run, not that it runs reliably.
Of the 4 api projects Argusic has installed and timed, one-api was the 3rd fastest to reach a running state, and 2 of 4 reached one at all.
What it is a reasonable choice for
- Trying it on a laptop or a small server: it reached a running state without a GPU.
Also tested, in the same area
Every one of these was installed and run by Argusic on a clean machine. Nothing appears here that was not tested.
Run history
Topics (from GitHub)
apiapi-gatewayazure-openai-apichatgptclaudeernie-botgeminigptopenaiopenai-apiproxy
Embed the badge
Markdown for the project README. It links back here; terms on the terms page.
[](https://argusic.com/subject/one-api)Questions
- Does one-api run?
- Yes. one-api runs. Argusic installed and launched it on a clean machine in 14 minutes, hitting 2 errors on the way, and recorded the session.
- How did Argusic test one-api?
- On a fresh, disposable machine, with every command recorded and the repository pinned at commit 8df4a2670b98. One attempt is recorded, and the full method is on the methodology page.
- What did it cost to test one-api?
- The run that produced this verdict cost $0.11: real compute and model cost, not a list price per million tokens. It is recorded on that run's page.
- How long does one-api take to install?
- 14 minutes in the run that decided this verdict, from a clean machine to a running state. That is one measured attempt, not an average.
- Does one-api need a GPU?
- Not to start. Argusic reached a running state on a machine with no GPU.
- What goes wrong when installing one-api?
- 2 things broke in the recorded run, and 2 were fixed without human help. Each one, and the time it cost, is listed on this page.
- How does one-api compare with the alternatives?
- Of the 4 api projects Argusic has installed and timed, one-api was the 3rd fastest to reach a running state, and 2 of 4 reached one at all.
- Where is the evidence for one-api?
- The recorded run is on this page, each linking to its full log and terminal recording, stored with a sha256 fingerprint so it cannot be quietly altered.