Use LibreChat with GammaInfra

Drop GammaInfra's smart router into LibreChat as a custom endpoint via the endpoints.custom block in librechat.yaml. Your self-hosted UI talks to every major LLM through one API key — with cost visibility and automatic provider fallback.

What changes with GammaInfra

• One custom endpoint covers everything. Instead of separate endpoint definitions for OpenAI, Anthropic, Google, Mistral, etc., add GammaInfra once. The full catalog populates the model dropdown.
• Per-request cost. LibreChat doesn't show cost natively. The GammaInfra dashboard rolls up exact per-request spend across users.
• Fallback on provider throttle. Your local UI doesn't break when one provider rate-limits — GammaInfra cascades transparently.

Setup

1. Get a GammaInfra API key

Sign up at gammainfra.com and verify your email.

2. Edit librechat.yaml

The configuration file lives in your LibreChat install root. If you don't have one yet, copy librechat.example.yaml to librechat.yaml and start there.

3. Add GammaInfra under endpoints.custom

endpoints:
  custom:
    - name: "GammaInfra"
      apiKey: "${KRAKEN_API_KEY}"
      baseURL: "https://api.gammainfra.com/v1"
      models:
        default:
          - "gammainfra/auto"
          - "gammainfra/fast"
          - "gammainfra/cheap"
          - "anthropic/claude-opus-4-7"
          - "anthropic/claude-sonnet-4-6"
          - "openai/gpt-5"
          - "openai/gpt-5-mini"
          - "google/gemini-3.1-pro-preview"
          - "deepseek/deepseek-v4-pro"
        fetch: false
      titleConvo: true
      titleModel: "gammainfra/cheap"
      modelDisplayLabel: "GammaInfra"

Save the file.

4. Set the env var

The ${KRAKEN_API_KEY} reference reads from your environment. In your LibreChat .env file:

KRAKEN_API_KEY=sk-gammainfra-...

5. Restart LibreChat

For Docker:

docker compose restart api

For direct installs:

pm2 restart librechat
# or whatever process manager you use

6. Verify in the UI

Open your LibreChat instance. The endpoint selector should now include "GammaInfra". Picking it shows the model list you configured under models.default.

Pricing display (optional)

LibreChat supports a tokens block under each custom endpoint to display token usage. Since GammaInfra passes through provider rates, you can set realistic per-model pricing for in-UI cost estimates:

endpoints:
  custom:
    - name: "GammaInfra"
      # ... (other fields)
      tokens:
        # Approximate — see /v1/models for actual provider rates
        "gammainfra/auto":               { input: 0.0015, output: 0.0060 }
        "anthropic/claude-opus-4-7": { input: 0.0150, output: 0.0750 }
        "openai/gpt-5-mini":         { input: 0.0003, output: 0.0024 }

Per-1k-token pricing. Refer to GET /v1/models for authoritative current rates; the GammaInfra dashboard reports exact billed cost regardless of what LibreChat estimates locally.

Trade-offs

• Latency. ~10–50 ms overhead per request. Imperceptible for chat-UI use.
• Cost. 3% top-up fee (launch) / 5% standard. Pass-through provider rates on tokens — no markup. BYOK 1–2% per request as an alternative.
• Privacy. GammaInfra doesn't log prompts or responses by default. Privacy policy. LibreChat still stores chat history in its own MongoDB per its config.

Troubleshoot

• Endpoint doesn't appear. YAML indent error. Make sure custom: is a list (dash prefix on the first field).
• "Authentication failed" on chat. KRAKEN_API_KEY env var not loaded — check the .env file lives in the right directory and the process re-read it on restart.
• Models missing from dropdown. Typo in a model name. Pull GET /v1/models for the authoritative list.

Ready to try it?

Get a GammaInfra API key →

$3 free trial credit on signup, $10 minimum top-up. Pass-through provider token rates plus 3% top-up fee during the launch window.