Ask the AI to help you build an EQL query.

The response message contains the AI's reply in Markdown. When it generates EQL, it will be in a JSON code block that you can extract, validate, and submit.

Use sessionId in follow-up requests to maintain conversation context.

Cost

API mode: 1 + dynamic credits

• base request charge: 1 credit
• dynamic usage charge: ceil(request_cost * 750) credits

x402 mode is quoted in USD per turn, not credits. The 402 response lists the schemes it accepts and the client picks one:

• exact (flat, charged in full): fast $1, expert $2
• batch-settlement (Base, when listed: deposit once into a payment channel; each turn is capped at fast $2, expert $6 and charged its actual cost)

Auth

Required header:

x-elfa-api-key: <api_key>

Chat produces drafts only; subsequent activation goes through POST /queries/drafts/{draftId}/convert.

Request Example

{
  "message": "Alert me when BTC breaks 100k",
  "speed": "expert",
  "sessionId": "optional-session-id"
}

Response Example (200)

{
  "sessionId": "session-uuid",
  "response": "I can help with that...",
  "title": "BTC Breakout Alert",
  "reasoning": null,
  "planIds": [],
  "credits": 104
}

credits is what this call cost, and carries the same total as the x-elfa-credits response header. Chat is dynamically priced, so read it rather than assuming a flat cost.

Best practice: persist sessionId and reuse it for follow-up prompts so the model keeps context across turns.