Plug in any LLM provider

Capture LLM token usage from any JSON. Automatically.

Forward your provider's raw response — OpenAI, Anthropic, Bedrock, Azure, Gemini, your custom proxy — and we'll detect prompt_tokens, total_tokens, model, cost & more. Per-tenant SQLite isolation, infinite-scroll dashboard, Excel export, full API.

Multi-tenant SQLite Auto schema detection Bring-your-own JSON

How it works

Four steps from signup to dashboard.

1. Register

2. POST your JSON

Forward provider responses to /api/v1/ingest/{customer}/{source}/{event} with your key.

3. Auto-detected

We find usage.prompt_tokens, input_tokens, model & cost — whatever shape the provider returns.

4. Explore

Infinite-scroll table, charts, Excel export. Edit any row, remap any field on the mapping screen.

A single API call

Same endpoint, any provider.

Pick a source and event name — each combination gets its own SQLite file so reads stay fast even at hundreds of millions of rows.

OpenAI, Anthropic, Azure OpenAI, Bedrock, Vertex, custom proxies
Per-tenant filesystem isolation
One ingest endpoint, one auth header

curl -X POST https://yourhost/api/v1/ingest/c_abc123/openai/chat \
  -H "X-Api-Key: ark_..." \
  -H "Content-Type: application/json" \
  -d '{
    "id": "chatcmpl-xyz",
    "model": "gpt-4o",
    "usage": {
      "prompt_tokens": 142,
      "completion_tokens": 87,
      "total_tokens": 229
    }
  }'