OmniRoute
App : A product you use directly
Local open-source gateway in front of 290+ providers with automatic quota-based switching.
Powerful, but aggregating free quotas and web cookie-based pools raise real security and terms-of-service concerns.
Security reflects where data goes, requested permissions and incidents on record at the verification date. It is not a guarantee of zero risk: check the deployment mode and the connections you enable.
What it does, concretely
Managing multiple AI providers in parallel usually means juggling separate API keys, different quota limits and incompatible SDKs. OmniRoute solves this by exposing a single local OpenAI-compatible endpoint, behind which it orchestrates 352 providers with automatic fallback when a provider is unavailable or out of quota.
In practice, you install OmniRoute via npm, Docker or as a desktop application, then point any OpenAI-compatible tool (Claude Code, Cursor, Cline, Codex, GitHub Copilot, etc.) at the local address http://localhost:20128/v1. A Next.js dashboard lets you configure routing strategies, monitor quotas in real time and manage encrypted credentials. A 12-engine compression pipeline reduces tokens sent by 15 to 95% on eligible content.
OmniRoute is aimed at developers who want to maximise free tiers across multiple providers, reduce API costs or avoid interruptions caused by rate limits, without changing their existing tooling. The catalogue documents roughly 1.47 billion free tokens per month by aggregating 34 documented recurring pools.
On the other hand, the tool is not suited to teams or individuals who do not want to manage a local gateway, who handle sensitive data without being able to audit the code, or who use ChatGPT, Claude or Gemini accounts via session cookies without having verified compatibility with those services' terms of use.
What it offers
- ·Automatic routing across 352 AI providers through a single local OpenAI-compatible endpoint, with fallback on unavailability or quota exhaustion.
- ·19 routing strategies available: load balancing, cost optimisation, context relay, 15-factor adaptive scoring and other modes configurable from the dashboard.
- ·12-engine compression pipeline (RTK, Caveman, Ultra, etc.) that reduces eligible tokens by 15 to 95%, with a documented average of around 89% on tool-heavy examples.
- ·Pool-deduplicated free-tier catalogue covering roughly 1.47 billion tokens per month from 34 documented recurring pools, with real-time tracking on the dashboard.
- ·Built-in MCP server exposing 110 tools across 33 scopes, and an A2A JSON-RPC server with 6 skills, usable without additional configuration.
- ·Persistent conversational memory combining FTS5 full-text search and vector recall via Qdrant.
- ·Semantic cache: repeated requests can be served from cache, with the result reported in the response header.
- ·Compatibility with 35 declared client integrations, including Claude Code, Codex, Cursor, Cline, GitHub Copilot, Gemini CLI, Aider, Goose and other OpenAI-compatible tools.
In practice
- Price
- Free
- Commercial use
- Commercial use: unclear
- Available on
- linux, macos, windows, cli, docker
- Licence
- MIT
Security and data
- Rating
- Security: serious reservations
- Your data
- Local or cloud depending on use
- Known incidents
- CVE-2026-49352 ; analyses pointant des contournements de CGU fournisseurs.
- Maintenance
- Active
- Last release
- 19 October 2026
Links and repository checked on 25 October 2026 · added on 1 February 2026
How to install OmniRoute
npm install -g omniroute