#4444th of 72 in Gateways and observability
octopus
LLM API gateway that aggregates provider channels behind one endpoint
- Stars
- 2.7k
- License
- AGPL-3.0
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- TypeScript
1,969 on 29 Apr 20262,749 on 18 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
Octopus is a self-hosted gateway written in Go with a bundled web panel. It groups upstream channels (OpenAI Chat, OpenAI Responses, Anthropic, Gemini) under one model name, converts between OpenAI and Anthropic API formats, and fails over when a channel errors. It also syncs model lists and prices from models.dev, and records request statistics, token use and cost.
Who it is for: Individuals routing several LLM provider keys through one endpoint
Strengths
- Converts between OpenAI Chat, OpenAI Responses and Anthropic formats; README shows Claude Code and Codex setup
- Single binary with embedded frontend; SQLite by default, MySQL or PostgreSQL optional
- Groups map several provider channels to one model name, with automatic failover
- Syncs model lists and pricing from models.dev, with manual price overrides
Weaknesses
- Statistics are batched in memory; kill -9 can lose them
- Default login is admin/admin and must be changed manually
- AGPL-3.0 license may restrict use in proprietary networked services
- Aimed at individuals; README describes no team or multi-user access control
What it needs
- no GPU
- Compose
- Needs SQLite (default), MySQL or PostgreSQL
- Models: OpenAI Chat, OpenAI Responses, Anthropic, Gemini
- port 8080
What changed
- v0.13.13 ↗ (opens in a new tab)Security fix
CVE-2026-111943. “🚨 CVE-2026-111943 - by @bestruirui (8aa58)”
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Known problems
Open now, reported in the last 90 days
Reported, then fixed by a linked change
- [Bug] 支持DeepSeek和Nvidia提供商启用禁用思考模式 ↗ (opens in a new tab)0 reactions · closed 5 May 2026Fixed by feat(transformer): add support for DeepSeek and Nvidia thinking mode ↗ (opens in a new tab)
- [Bug] claude-opus-4-6 在新版 cc 中经 octopus 调用时 Thinking 启用无效;切回 claude-opus-4-5 正常 ↗ (opens in a new tab)0 reactions · closed 3 Mar 2026Fixed by fix: 修复 Anthropic Adaptive Thinking 支持及 BudgetTokens 空指针问题 ↗ (opens in a new tab)
- [Bug] 更新火山引擎 supportedReasoningEffortModel ↗ (opens in a new tab)0 reactions · closed 25 Jan 2026Fixed by fix: update supportedReasoningEffortModel close #132 ↗ (opens in a new tab)
- [Bug] Cherry Studio 请求异常 ↗ (opens in a new tab)0 reactions · closed 23 Jan 2026Fixed by fix: handle missing choices field in stream response ↗ (opens in a new tab)
Titles are the reporters' own words, taken from the project's GitHub issues and ranked by reactions. We have not tried to reproduce any of them.
Also in Gateways and observability
See all 72| Rank | Project | Score |
|---|---|---|
| 1 | OmniRouteLLM gateway routing one OpenAI-compatible endpoint across 372 providers | 96 out of 100 |
| 2 | headroomCompresses tool outputs, logs and history before they reach the LLM | 92 out of 100 |
| 3 | MLflowTracing, evaluation and prompt registry for LLM apps, plus ML tracking | 92 out of 100 |
| #4 | new-apiGateway that unifies LLM providers behind OpenAI, Anthropic and Gemini APIs | 91 out of 100 |
| #5 | LiteLLMOpenAI-format gateway and Python SDK for calling 100+ LLM providers | 88 out of 100 |