OPENAI SDK COMPATIBLE

Eachlabs LLM Router

Use the OpenAI SDK you already know. Just change the base URL to call our featured models, or any chat model ID listed on OpenRouter, through a single, intelligent router layer.

// Just change the baseURL
baseURL: 'https://api.eachlabs.ai/v1'

Quick Start

Get started in under 2 minutes with your favorite language.

  • npm i openai

Featured models

Our curated picks, shown with the model ID to send. Chat Completions also accepts any other chat model ID listed on OpenRouter.

  • openai/gpt-5.6-sol

    OpenAI flagship GPT-5.6 model for complex reasoning, coding, agentic workflows, and long-horizon problem solving through the Eachlabs LLM Router.

    When should I use GPT-5.6 Sol?

    Use GPT-5.6 Sol for demanding reasoning, coding, and multi-step agentic workflows that need the strongest GPT-5.6 model available in the router.

    How is GPT-5.6 Sol routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • openai/gpt-5.6-terra

    Balanced OpenAI GPT-5.6 model for everyday coding, reasoning, and agentic tasks through the Eachlabs LLM Router.

    When should I use GPT-5.6 Terra?

    Use GPT-5.6 Terra when you want a balance of GPT-5.6 capability and cost for coding, reasoning, and general assistant workflows.

    How is GPT-5.6 Terra routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • openai/gpt-5.6-luna

    Fast, cost-efficient OpenAI GPT-5.6 model for high-volume chat, classification, lightweight reasoning, and latency-sensitive workflows through the Eachlabs LLM Router.

    When should I use GPT-5.6 Luna?

    Use GPT-5.6 Luna for high-volume or latency-sensitive workloads that need capable reasoning at a lower price tier.

    How is GPT-5.6 Luna routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • openai/gpt-5.2

    OpenAI frontier model for advanced reasoning, coding, tool use, and production text generation through the Eachlabs LLM Router.

    When should I use GPT-5.2?

    Use GPT-5.2 for high-quality reasoning, code generation, long-form writing, and complex tool workflows.

    How is GPT-5.2 routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • openai/gpt-5.5

    OpenAI frontier model for advanced reasoning, coding, and high-quality assistant workloads through the Eachlabs LLM Router.

    When should I use GPT-5.5?

    Use GPT-5.5 for demanding reasoning, coding, and production assistant workloads that need the strongest OpenAI model available in the router.

    How is GPT-5.5 routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • typesafe/jev-1.13

    TypeSafe Jev 1.13 evaluates native noul, choice, and score questions through the Decisions API. Availability is gated per environment; chat completions and Responses are unsupported.

    How do I call Jev 1.13?

    When enabled, send POST /v1/decisions with model typesafe/jev-1.13, state, and a nonempty questions object. See https://docs.eachlabs.ai/llm-router/decisions. Do not use chat completions or Responses.

    When should I use Jev 1.13?

    Use Jev 1.13 for fast routing, classification, and structured decision tasks that need a typed choice instead of long-form generation.

  • openai/gpt-5.5:online

    OpenAI frontier model for advanced reasoning, coding, and high-quality assistant workloads through the Eachlabs LLM Router.

    When should I use GPT-5.5?

    Use GPT-5.5 for demanding reasoning, coding, and production assistant workloads that need the strongest OpenAI model available in the router.

    How is GPT-5.5 routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • openai/gpt-5.4-mini

    Efficient OpenAI model for lower-latency chat, extraction, classification, and lightweight reasoning through the Eachlabs LLM Router.

    When should I use GPT-5.4 Mini?

    Use GPT-5.4 Mini when you want an OpenAI model optimized for speed and cost while retaining broad general-purpose capability.

    How is GPT-5.4 Mini routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • openai/gpt-5.4-mini:online

    Efficient OpenAI model for lower-latency chat, extraction, classification, and lightweight reasoning through the Eachlabs LLM Router.

    When should I use GPT-5.4 Mini?

    Use GPT-5.4 Mini when you want an OpenAI model optimized for speed and cost while retaining broad general-purpose capability.

    How is GPT-5.4 Mini routed?

    The router resolves this id to a qualified upstream provider; usage is billed on your Eachlabs account.

  • openai/gpt-5.1

    OpenAI frontier model for balanced reasoning, speed, and general-purpose assistant workloads through the Eachlabs LLM Router.

    When should I use GPT-5.1?

    Use GPT-5.1 when you need a strong general-purpose model with dependable reasoning and broad task coverage.

    Is GPT-5.1 the canonical model id?

    Yes. The catalog uses the canonical model id and keeps legacy aliases hidden.

  • openai/gpt-4.1

    OpenAI model for reliable text generation, extraction, coding, and structured assistant workflows.

    When should I use GPT-4.1?

    Use GPT-4.1 for dependable general-purpose generation and structured output workloads.

    Does this catalog entry change routing behavior?

    No. This entry documents the existing explicit route and does not alter runtime routing.

  • openai/gpt-4.1:online

    OpenAI model for reliable text generation, extraction, coding, and structured assistant workflows.

    When should I use GPT-4.1?

    Use GPT-4.1 for dependable general-purpose generation and structured output workloads.

    Does this catalog entry change routing behavior?

    No. This entry documents the existing explicit route and does not alter runtime routing.

Audio endpoints

OpenAI-compatible speech-to-text and text-to-speech routes that are callable right now. Each entry below is confirmed available by the API; models listed as unavailable are not reachable on these endpoints.

Speech to text

/v1/audio/transcriptions
  • Whisper Large v3

    Available

    openai/whisper-large-v3

    Bounded speech-to-text through the OpenAI-compatible audio transcription endpoint.

    Languages
    Auto or iso code
    Input formats
    flac, m4a, mp3, mp4, mpeg, mpga, ogg, wav, webm
    Response formats
    json, verbose_json
    Timestamps
    word, segment
    Max upload
    25.0 MB
    Read the docs
  • Whisper Large v3 Turbo

    Available

    openai/whisper-large-v3-turbo

    Fast bounded speech-to-text through the OpenAI-compatible audio transcription endpoint.

    Languages
    Auto or iso code
    Input formats
    flac, m4a, mp3, mp4, mpeg, mpga, ogg, wav, webm
    Response formats
    json, verbose_json
    Timestamps
    word, segment
    Max upload
    25.0 MB
    Read the docs
  • xAI Speech to Text

    Available

    xai-speech-to-text

    Bounded xAI transcription with OpenAI-compatible JSON and word timestamps.

    Languages
    Auto or iso code
    Input formats
    flac, m4a, mp3, mp4, mpeg, mpga, ogg, wav, webm
    Response formats
    json, verbose_json
    Timestamps
    word
    Max upload
    25.0 MB
    Read the docs
  • Deepgram Nova-3 Speech to Text Pro

    Available

    deepgram-nova-3-speech-to-text-pro

    Bounded Deepgram Nova-3 transcription with OpenAI-compatible word and segment timestamps.

    Languages
    Auto or iso code
    Input formats
    flac, m4a, mp3, mp4, mpeg, mpga, ogg, wav, webm
    Response formats
    json, verbose_json
    Timestamps
    word, segment
    Max upload
    25.0 MB
    Read the docs
  • Muse Voice Transcribe

    Available

    meta-muse-voice-transcribe

    Bounded Muse Voice Transcribe file transcription with turn-level timestamps.

    Languages
    Auto or iso code
    Input formats
    flac, m4a, mp3, mp4, mpeg, mpga, ogg, wav, webm
    Response formats
    json, verbose_json
    Timestamps
    segment
    Max upload
    25.0 MB
    Read the docs

Text to speech

/v1/audio/speech
  • Deepgram Aura-2

    Available

    deepgram/aura-2

    Progressively streamed English text-to-speech through the OpenAI-compatible audio speech endpoint.

    Languages
    English only
    Voices
    aura-2-thalia-en
    Response formats
    mp3, pcm
    Max input
    2,000 characters
    Read the docs
  • Gemini 3.1 Flash TTS Preview

    Available

    google/gemini-3.1-flash-tts-preview

    Progressively streamed multilingual text-to-speech with PCM output through the OpenAI-compatible audio speech endpoint.

    Languages
    Auto or iso code
    Voices
    Zephyr, Puck, Kore, Charon, Aoede
    Response formats
    pcm
    Max input
    5,000 characters
    Read the docs
  • Gemini 3.1 Flash Text to Speech

    Available

    gemini-3-1-flash-text-to-speech

    Progressively streamed multilingual text-to-speech through the OpenAI-compatible audio speech endpoint.

    Languages
    Auto or iso code
    Voices
    Zephyr, Puck, Kore, Charon, Aoede
    Response formats
    pcm
    Max input
    5,000 characters
    Read the docs
  • xAI Grok Text to Speech

    Available

    xai-grok-tts-text-to-speech

    Progressively streamed xAI text-to-speech with automatic language detection.

    Languages
    Automatic
    Voices
    ara, eve, leo, rex, sal
    Response formats
    mp3, wav, pcm
    Speed
    0.7×–1.5×
    Max input
    15,000 characters
    Read the docs
  • MiniMax Speech 2.8 Turbo

    Available

    minimax-speech-2-8-turbo

    Progressively streamed MiniMax Speech 2.8 Turbo text-to-speech using reviewed English system voices.

    Languages
    English only
    Voices
    English_expressive_narrator, English_radiant_girl, English_magnetic_voiced_man
    Response formats
    mp3
    Speed
    0.5×–2×
    Max input
    9,999 characters
    Read the docs

Not available on these routes

  • Google Text to Speechgoogle-text-to-speechUnavailable

    This resolves to the same upstream as gemini-3-1-flash-text-to-speech, which is enabled; use that id rather than maintaining two names for one model.

  • Alibaba Qwen3 ASR Flash FileTransalibaba-qwen3-asr-flash-filetrans-speech-to-textUnavailable

    This model is asynchronous upstream: audio is submitted as a job and the transcript is retrieved by polling, which the single-response transcription endpoint cannot express. It remains available through prediction and polling.

  • ElevenLabs Speech to Textelevenlabs-speech-to-textUnavailable

    Not yet available on the OpenAI-compatible audio transcription endpoint while privacy and billing qualification is completed.

Keep your code. Change the base URL.

  • Zero Code Changes

    Your existing OpenAI SDK code works instantly. Just swap the base URL and API key.

  • Intelligent Routing

    Our LLM router automatically selects the best model provider for your request.

  • Pay-as-you-go

    No monthly fees. Only pay for what you use with transparent per-request pricing. You pay the model's list price plus the same standard platform fee OpenRouter charges — nothing more.

Frequently asked questions

The short answers. For anything specific to your pipeline, ask the people who build the platform.

Talk with an engineerFind your next pipeline opportunity.
  • each::labs features a curated set of frontier and open-weight LLMs, including GPT, Claude, Gemini, Grok, Llama, Mistral, Qwen, and DeepSeek. The OpenAI-compatible LLM Router also accepts any other chat model ID listed on OpenRouter, except free (:free) variants, through the same endpoint and API key.

Ready to get started?

Create your API key and start building in minutes.