Bankr Dev - API Basics

Bankr Agent API Essentials

When answering questions about the Bankr API:

1. Explain the asynchronous job pattern (submit-poll-complete)
2. Reference the endpoint documentation and TypeScript examples below
3. For detailed response structures, consult references/job-response-schema.md
4. Point developers to working examples in examples/ directory

---

The Bankr Agent API enables programmatic access to crypto trading, market analysis, and prediction markets through a simple asynchronous job pattern.

Core Concept: Asynchronous Job Pattern

All Bankr operations follow a submit-poll-complete pattern:

1. Submit a natural language prompt to start a job
2. Poll the job status every 2 seconds
3. Receive results when status is terminal (completed/failed/cancelled)

This pattern handles operations that may take 30 seconds to 2+ minutes (trades, complex analysis).

API Endpoints

Base URL

https://api.bankr.bot

Authentication

All requests require the x-api-key header:

x-api-key: bk_your_api_key_here

The API key is tied to a specific user's Bankr account and wallet.

Endpoint 1: Submit Prompt

POST /agent/prompt
Content-Type: application/json

{
  "prompt": "Buy $50 of ETH on Base"
}

Response:

{
  "success": true,
  "jobId": "job_abc123",
  "status": "pending"
}

Code example:

async function submitPrompt(prompt: string): Promise<{ jobId: string }> {
  const response = await fetch(`${API_URL}/agent/prompt`, {
    method: "POST",
    headers: {
      "x-api-key": API_KEY,
      "Content-Type": "application/json",
    },
    body: JSON.stringify({ prompt }),
  });
  const data = await response.json();
  if (!data.success) throw new Error(data.error || "Failed to submit");
  return { jobId: data.jobId };
}

Endpoint 2: Get Job Status

GET /agent/job/{jobId}

Response:

{
  "success": true,
  "jobId": "job_abc123",
  "status": "completed",
  "prompt": "What is the price of ETH?",
  "response": "Ethereum (ETH) is currently trading at $3,245.67...",
  "transactions": [],
  "statusUpdates": [
    { "message": "Fetching price data...", "timestamp": "2024-01-15T10:00:02Z" }
  ],
  "createdAt": "2024-01-15T10:00:00Z",
  "completedAt": "2024-01-15T10:00:05Z",
  "processingTime": 5000
}