OmniRoute has joined Cheaper InferenceExplore Cheaper Inference

Никога не спирай да кодираш.

Route across 352 providers through one OpenAI-compatible endpoint, with automatic fallback.

OpenAI-compatible Free and open source61.1k stars (opens in a new tab)

OmniRoute × Cheaper Inference

Combine OmniRoute’s automatic fallback and developer tools with Cheaper Inference’s cost-ranked marketplace. Every eligible request starts with the cheapest available provider route.

63 modelsDirect list price compared with the Cheaper Inference price
Direct list prices compared with Cheaper Inference prices by model
ModelInput priceOutput priceSavingsContextDetails
grok-imaginexAI · image$0.01$0.0014/ image$0.05$0.014/ image72% off
gpt-5.6-lunaOpenAI · text$0.20$0.08/ 1M tokens$1.20$0.48/ 1M tokens60% off1.1M
gpt-5.6-terraOpenAI · text$2.00$0.80/ 1M tokens$12.00$4.80/ 1M tokens60% off1.1M
gpt-oss-120bOpenAI · text$0.10$0.04/ 1M tokens$0.50$0.20/ 1M tokens60% off
gpt-5.6-solOpenAI · text$2.00$1.00/ 1M tokens$10.00$5.00/ 1M tokens50% off1.1M
glm-5Z.ai · text$1.00$0.534597/ 1M tokens$3.20$1.71071/ 1M tokens46.54% off198K
glm-4.5Z.ai · text$0.60$0.33/ 1M tokens$2.20$1.21/ 1M tokens45% off128K
glm-4.5-airZ.ai · text$0.20$0.11/ 1M tokens$1.10$0.605/ 1M tokens45% off128K
Showing 8 of 63 models

Обещанието

Един endpoint. Никога не спирай да изграждаш.

Core control

Никога не достигай лимити

Автоматичен fallback сред 339 доставчика за милисекунди. Квотата свърши? Следващият доставчик поема. нулев престой.

Core control

Спести до 95% токени

A 12-engine pipeline reduces eligible context by 15–95%. The documented tool-heavy example reduced it by ~89%.

Core control

$0 за начало

90+ free tiers and 56 recurring or keyless free-forever providers. No credit card needed for the keyless quick start.

Core control

35 supported integrations

16+ агента за кодиране. Claude Code, Codex, Cursor, Cline, Copilot, Antigravity. чрез една конфигурация.

Core control

Един endpoint

Превод OpenAI ↔ Claude ↔ Gemini ↔ Responses API. Насочи всеки инструмент към /v1 и просто работи.

Core control

Готов за production

Прекъсвачи, TLS stealth, MCP (105 инструмента), A2A, памет, ограничители, оценки. подкрепено с 25,000+ теста.

Каталогът

All behind one endpoint.

A catalog spanning chat, media, search, local runtimes, cloud agents and system integrations.

352 providers

Every one of them, behind a single local endpoint.

The one endpoint every client points athttp://localhost:20128/v1

14 Search12 Audio-only2 Upstream proxy3 Cloud agent1 System
OpenAIAnthropicGoogleDeepSeekxAIZ.aiMoonshotMiniMaxAlibabaMetaBytedanceAion Labs
25OAuthClaude Code, Codex, Cursor, GitHub Copilot, Kimi Code…
233API keyOpenAI, Anthropic, Groq, NVIDIA, Cerebras, Mistral…
35Web cookieChatGPT Web, Claude Web, Gemini Web, Grok Web…
14LocalOllama, LM Studio, vLLM, llama.cpp…
13No authenticationOpenCode Free, Felo, Pollinations and other keyless routes…
Explore the provider reference (opens in a new tab)

Работещ за 60 секунди

Три команди. Нула конфигурация.

Install the package, launch the gateway, then point any OpenAI-compatible client at port 20128.

Provider catalog
352
Routing strategies
19
Local endpoint
:20128/v1

Step 1: Install

Install the omniroute npm package globally.

omniroute: install
$ npm install -g omnirouteadded 1 package

Step 2: Run

Launch the API and dashboard together on port 20128.

omniroute: run
$ omniroute▸ dashboard ✓ http://localhost:20128/dashboard▸ api ...... ✓ serving on :20128

Step 3: Connect your tool

Use http://localhost:20128/v1; the keyless auto route works on a fresh install.

omniroute: connect
$ curl localhost:20128/v1/models✓ models listed

Всичко включено

Всичко вградено. Без добавки.

Маршрутизиране, устойчивост, изчерпване на безплатни квоти, компресия и агентски протоколи. един бинарен файл, нула плъгини за свързване.

Gateway control

Умни комбинации

Set a model to auto or build your own combo with 19 strategies and tiered fallback.

Use a ready-made mode or combine 19 strategies around the goal that matters.

Gateway control

Памет

Постоянна разговорна памет. FTS5 ключови думи + Qdrant векторно извикване, вградено.

Gateway control

3-layer resilience

Провал на правилното ниво. никога на цялото. Circuit breaker на доставчик, cooldown на връзка, lockout на модел. Другите ключове и модели продължават да обслужват.

Gateway control

Безплатна квота

Pool-deduped accounting across 90+ free tiers.

~1.51B безплатни токена на месец.

Gateway control

Компресия

12 composable engines for tool output, context and response style.

Reduce eligible tokens by 15–95%.

Gateway control

MCP Server

A built-in server exposing 110 gateway tools across 33 scopes.

110 tools · 33 scopes · stdio / HTTP

Gateway control

A2A протокол

A JSON-RPC agent server with 6 skills.

6 skills · JSON-RPC 2.0 · Agent Card

Gateway control

Semantic cache

A repeated request can return from cache, and the response header reports what that saved.

X-OmniRoute-Cache: HIT

Client compatibility

Keep your existing coding tools.

Point an OpenAI-compatible CLI or editor at one local base URL; OmniRoute handles the providers behind it.

Claude Code
Codex CLI
Cursor
Cline
GitHub Copilot
Gemini CLI
OpenCode
Kilo Code
Factory Droid
Continue
Roo Code
Antigravity
Aider
OpenClaw
Goose
Grok Build
Windsurf
Devin CLI

http://localhost:20128/v1

request.sh
curl http://localhost:20128/v1/chat/completions \
  -H "Authorization: Bearer $OMNIROUTE_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"auto","messages":[{"role":"user","content":"Hello"}]}'

Ново

Вече във вашия VS Code Copilot Chat

Инсталирайте разширението OmniCopilot и всеки модел на OmniRoute се появява директно в селектора на модели, който вече използвате. без нов страничен панел, без нов чат интерфейс.

Official Marketplace

Works in VS Code stable and Insiders. Install from the Marketplace or Extensions view.

Open VSX Registry

За форкове на VS Code, които не използват официалния магазин. Cursor, Windsurf, VSCodium, Theia, code-server, Gitpod и други.

Не е нужен план за Copilot

От VS Code 1.122 моделите на доставчиците работят без вход в GitHub или абонамент за Copilot. агентен режим, извикване на инструменти и визия, върху безплатните модели на OmniRoute.

Ежедневната реалност

Защо OmniRoute

Всеки проблем с coding-AI има вграден отговор. Без скриптове, без наблюдение.

Without routing
With OmniRoute
Quota expires unused
Track and drain the allowance you already have
Rate limits stop you mid-task
4-tier auto-fallback in milliseconds
Tool output burns tokens
12-engine compression, 15–95% on eligible content
Expensive APIs add up
Cost-optimized routing can pick the cheapest eligible target
Each tool needs its own setup
One endpoint and one dashboard
A provider is unavailable from your network
Proxy scopes and client-fingerprint controls
One dead key breaks a provider
Breaker, cooldown and lockout isolate the failure
A team shares one subscription
Key pools and fair-share quotas
You need local control of routing data
Local-first storage with encrypted credentials
No spend or health visibility
Usage, quota, savings and latency analytics

Честно сравнение

Как се сравняваме

Capabilities checked against the linked project documentation on 2026-08-30.

CapabilityOmniRoute (opens in a new tab)9Router (opens in a new tab)LiteLLM (opens in a new tab)CLIProxyAPI (opens in a new tab)
Providers35240+100+8+ upstreams
Routing strategies193-tierretry / priorityround-robin / fill-first
Tiered fallback with UIYesYesManual policyMulti-account
Token compression12 engines · 15–95%RTK / Headroom / CavemanNo built-in pipelineHTTP compression

Public README and official documentation snapshot; verify linked sources before purchase or deployment decisions.

Routing

Choose an auto route.

Use a ready-made mode or combine 19 strategies around the goal that matters.

Model · fallback enabledautoBalanced default
OpenAI-compatible request
curl http://localhost:20128/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"auto","messages":[{"role":"user","content":"Write a TypeScript retry helper."}]}'

Routing alias auto selected. Objective: Balanced default. Fallback enabled.

Drain subscription quota

Use quota you already have before touching paid APIs.

priority · fill-first

Spread load

Balance traffic across keys and accounts to reduce rate-limit pressure.

weighted · round-robin · p2c · least-used · headroom

Cheapest first

Route to the lowest-cost eligible target.

cost-optimized

Context-aware

Keep long contexts on models that fit, then relay when needed.

context-relay · context-optimized · cache-optimized

Randomized

Distribute eligible traffic across targets.

random · strict-random

Adaptive scoring

Use 15-factor scoring, last-known-good and reset-aware windows.

reset-window · reset-aware · lkgp · auto · fusion · pipeline

Безплатни токени, честно преброени

~1.51B безплатни токена на месец.

455 catalog entries are deduplicated into 40 recurring pools; only published positive budgets feed the headline.

1.51B

Pool-deduped accounting keeps shared quotas from being counted more than once. No-cap providers stay visible without inflating the token total. Checked 2026-08-30.

View the calculation (opens in a new tab)
Recurring or keyless free-forever providers56
  • Mistral1B/month
  • LLM7150M/month
  • Nara150M/month
  • Gemini CLI · forever60M/month
  • Kiro50 credits/month
  • Qoder · foreverFree catalog
  • LongCat · forever50M/day
  • Cerebras1M/day
  • NVIDIA NIM~40 RPM
  • Pollinations · foreverNo key
  • Cloudflare Workers AI · foreverDaily neurons
  • OpenCode Free · foreverNo-auth catalog

Спри да изгаряш токени

Reduce eligible tokens by 15–95%.

Изходът на инструментите, повтарящият се контекст и раздутият JSON се компресират преди да достигнат модела. същото значение, малка част от токените. 12 композируеми двигателя се наслагват в конвейер, който вие контролирате.

Context reduced from 15–95% on eligible tokens

Before: The function should iterate over the list of items, and for each item, it should check whether the value is greater than zero, and if so, add it to the running total. After: sum all list items where value > 0

  1. Session-Dedup
  2. CCR
  3. Lite
  4. RTK
  5. Responses Tool Output
  6. Headroom
  7. Relevance
  8. Caveman
  9. Aggressive
  10. LLMLingua-2
  11. Ultra
  12. OmniGlyph

Same instruction. Less context to pay for.

Before69 tokens
The function should iterate over the list of items, and for each item, it should check whether the value is greater than zero, and if so, add it to the running total.
After19 tokens
sum all list items where value > 0

Composable pipeline

  1. Session-Dedup
    Drop content already sent earlier in the session.
  2. CCR
    Archive large blocks and restore them on demand.
  3. Lite
    Apply conservative lossless cleanup.
  4. RTK
    Filter and deduplicate command and tool output.
  5. Responses Tool Output
    Compact Responses API tool payloads.
  6. Headroom
    Apply reversible tabular and structured-data compaction.
  7. Relevance
    Keep sentences most relevant to the latest query.
  8. Caveman
    Use rule-based prose condensation.
  9. Aggressive
    Apply stronger configured pruning.
  10. LLMLingua-2
    Use code-safe semantic pruning in an isolated worker.
  11. Ultra
    Combine high-savings stages with preservation guards.
  12. OmniGlyph
    Encode eligible repeated structures compactly.

Profiles

Published profile range~30%x-omniroute-compression: standard

Standard compression selected. Published range ~30%.

Controls and output styles

  • Named profiles and active selector
  • Visual Compression Studio
  • Anthropic Context Editing
  • Adaptive context-budget dial
  • Per-request `x-omniroute-compression` header
  • Offline eval harness
  • Token-threshold auto-trigger
  • Terse prose
  • Less code (YAGNI)
  • Terse CJK 文言

Control precedence: request header → combo → profile → adaptive policy → default.

3 слоя, не 3 нива

Провал на правилното ниво. не на цялото.

Лошият ключ не трябва да унищожи доставчика. Мъртъв модел не трябва да унищожи ключ. Всеки провал е изолиран до най-малкия възможен обхват, така че всичко останало продължава да обслужва.

model ⊂ connection ⊂ provider
Provider circuit breakerWhen a provider is unhealthy, combo routing skips it until a half-open probe succeeds.
Connection cooldownCool down one key or account while sibling connections keep serving.
Model lockoutLock one rate-limited or unavailable model while other models on the connection remain available.

Network controls

Network routing and local data controls.

Proxy scopes control network routing while local-first storage keeps operational data on your machine.

Three proxy scopes

Route globally, per provider or per individual connection.

  • Global · per-provider · per-connection
  • SOCKS5 and HTTP(S) proxies
  • Proxy marketplace integration

TLS and client fingerprint controls

Match supported provider clients down to selected transport characteristics.

  • JA3/JA4-capable transport via wreq-js
  • Provider-specific client profiles
  • IPv4 and IPv6 egress controls

Private and local-first

Keys, usage and history stay on the host you control.

  • SQLite on disk with AES-256-GCM encrypted credentials
  • No OmniRoute account required
  • Self-host on infrastructure you choose
RussiaChinaIranCubaTürkiye

Виж всичко с един поглед

Пълно табло, не конфигурационен файл.

Manage providers, combos, analytics, health and costs at localhost:20128/dashboard.

localhost:20128/dashboardExample data
352providers connected
90+free tiers available
198healthy in this example
  • Claude Code · OAuth · live
  • Cerebras · free · live
  • Groq · cooldown example
  • NVIDIA NIM · free · live

Агент-ориентиран

Свържи агент. той управлява OmniRoute.

OmniRoute не е само прокси. той говори агентските протоколи. Включи го в MCP клиент, A2A мрежа или облачен coding агент и той се превръща в инструмент, с който агентът може да работи.

MCP сървър

110 tools across 33 scopes over stdio and HTTP transports.

110 tools · 33 scopes · stdio / HTTP

claude mcp add omniroute --type http --url http://localhost:20128/api/mcp/stream

A2A сървър

6 skills for smart routing, quota, discovery, cost analysis and health reporting over JSON-RPC 2.0.

6 skills · JSON-RPC 2.0 · Agent Card

Облачни агенти

Управлявай Codex, Cursor, Devin и Jules чрез един интерфейс. създавай задачи, одобрявай планове и предавай резултати поточно, всичко маршрутизирано чрез OmniRoute.

Codex · Cursor · Devin · Jules

Отвъд маршрутизирането

Не просто прокси. платформа.

Вградени услуги

Стартирай Bifrost (Go relay), 9Router или CLIProxy като управлявани sidecar-и. или пълен клъстерен профил. директно от таблото.

Твоето знание, в контекст

Notion и Obsidian свързани като MCP инструменти. твоите бележки и трезори стават първокласен контекст за всеки модел.

Вградени предпазни мерки

Защита от prompt-injection на всеки маршрут, опционална редакция на PII и нормализация на филтъра за безопасно съдържание. защита без обвързване с доставчик.

Геймификация

Серии, постижения и спестявания на живо. твоето използване, направено осезаемо.

From the community

How OmniRoute is used and discussed.

See what developers are building, asking and sharing across X, Reddit and YouTube.

32 videos from 32 channelsSee all 32 community videos

one local endpoint for cursor / cline / claude code / copilot 😳 OmniRoute (24.7k★ on GitHub) runs a self-hosted router on your machine so you only need one base URL for all your coding agents. wha…

25 posts from X · 25 Reddit postsSee every post

Изградено открито

Бързо развиваща се, активна общност

OmniRoute се обновява постоянно. движен от стотици сътрудници, които го държат напред.

OmniRoute by Cheaper Inference

Choose how you run OmniRoute.

Self-host OmniRoute, or start with Cheaper Inference for a hosted cost-ranked gateway.