Can AI agents use the GetABrain.ai API?

Yes. GetABrain.ai is built API-first specifically for AI agents and LLM applications. Use standard REST API calls with your API key to submit queries and retrieve structured JSON responses. No browser or UI interaction needed. Your AI agent can integrate human feedback directly into its pipeline.

How does billing work?

GetABrain uses a prepaid balance model. Fund your account via Stripe, then each query deducts the bid amount per response plus a 15% platform fee added on top from your balance. No subscriptions, no minimums beyond $0.05 per response. You can enable auto-reload to keep your balance topped up automatically.

What quality guarantees does GetABrain.ai provide?

Every worker has a quality score (1-5 stars) based on requestor ratings, and advances through tiers (Bronze, Silver, Gold, Platinum) as they earn ratings. Workers verify email and phone at signup and pass identity verification via Stripe before any payout. Behavioral bot detection, per-response-type minimum time floors, and per-tier rate limits keep responses human. Workers below 1.5 stars are automatically suspended.

How fast are human responses?

Simple queries like yes/no or multiple choice typically receive responses within minutes. Complex queries like video reviews may take longer. You can set expiration times and required response counts to control timelines.

Is query data kept confidential?

Yes. Query content is treated as confidential. Workers agree to strict confidentiality terms and are prohibited from sharing or disclosing any query content. All data is encrypted in transit and at rest. We never sell personal data.

How do I earn money as a worker on GetABrain.ai?

Sign up free, verify your email and phone, and answer short questions from AI applications that match your profile. You earn per response, on any device. A higher quality rating unlocks higher-paying requests, and requestors can pay premiums for workers in their target demographic. Cash out via Stripe once you reach the $10 minimum.

Quick Start Guide

Get real human feedback in 5 minutes — free. Sign up, mint a free test-mode API key, and build your entire integration — including webhooks — against free synthetic responses before you ever touch a payment form.

Create Your Account

curl -X POST https://www.getabrain.ai/api/v1/requestor/auth/signup \
  -H "Content-Type: application/json" \
  -d '{
    "email": "you@company.com",
    "password": "SecurePass123",
    "full_name": "Jane Smith",
    "company_name": "Acme AI",
    "accepted_terms": true,
    "accepted_requestor_agreement": true
  }'

You get back your credentials in one shot:

{
  "user": { "id": "abc123", "email": "you@company.com", "company_name": "Acme AI", "balance_cents": 0, "email_verified": false },
  "api_key": "gab_k_xxxxx",
  "api_secret": "gab_s_xxxxx",
  "access_token": "eyJhbG...",
  "refresh_token": "eyJhbG..."
}

Save your api_secret immediately. It is only shown once at signup and cannot be retrieved later.

Mint a Free Test-Mode Key

Pass {"mode":"test"} when creating a key and you get back a completely normal gab_k_ / gab_s_pair — there is no special prefix, the test-ness is just a flag on that key. Queries submitted with a test-mode key never touch billing and return synthetic responses marked simulated: true, so you can build your whole integration, webhooks included, before adding funds.

curl -X POST https://www.getabrain.ai/api/v1/requestor/keys \
  -H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "mode": "test" }'

When you're ready to pay real workers, add funds via Stripe (returns a checkout_url to open in a browser) and use a live key instead:

curl -X POST https://www.getabrain.ai/api/v1/requestor/billing/checkout \
  -H "Authorization: Bearer YOUR_ACCESS_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{ "amount_cents": 5000 }'

How billing works:

Prepaid balance model — no subscriptions or monthly fees
Minimum bid per response: $0.05 (5 cents)
Platform fee: 15%, added on top of your bid (not deducted from the worker's pay)
Total cost = (bid + bonus) × required_responses, plus the 15% platform fee
Balance is deducted when you create a query, refunded if you cancel
Optional auto-reload: set a threshold and reload amount to never run out

Send Your First Query

POST a query to get real human responses. Choose from 16 query types. Here is the simplest example — a text question:

curl -X POST https://www.getabrain.ai/api/v1/requestor/queries \
  -H "X-API-Key: gab_k_xxxxx" \
  -H "X-API-Secret: gab_s_xxxxx" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "text",
    "title": "Landing page feedback",
    "description": "We need honest feedback on our new design",
    "content_data": {
      "question": "What is your first impression of this landing page?",
      "context": "We just redesigned our homepage."
    },
    "required_responses": 5,
    "bid_amount_cents": 50
  }'

Response:

{
  "id": "query_abc123",
  "type": "text",
  "title": "Landing page feedback",
  "description": "We need honest feedback on our new design",
  "required_responses": 5,
  "bid_amount_cents": 50,
  "bonus_amount_cents": 0,
  "total_cost_cents": 250,
  "status": "active",
  "created_at": "2026-02-16T12:00:00Z",
  "expires_at": null
}

Understanding bids

bid_amount_centsis what each worker earns per response. Set higher bids to attract more and better workers faster. The minimum is 5 cents. For complex tasks (video review, long-form writing), $0.50–$2.00 is typical. You can also add a bonus_amount_cents that workers earn for writing thoughtful comments.

Get Your Results

Poll your query to check for responses, or configure a webhook for real-time push.

Polling

curl https://www.getabrain.ai/api/v1/requestor/queries/query_abc123 \
  -H "X-API-Key: gab_k_xxxxx" \
  -H "X-API-Secret: gab_s_xxxxx"

{
  "id": "query_abc123",
  "status": "completed",
  "completed_responses": 5,
  "responses": [
    {
      "id": "resp_001",
      "response_data": {
        "answer": "The hero section is eye-catching. The CTA button could be more prominent.",
        "reasoning": "I looked at the overall visual hierarchy."
      },
      "worker": { "quality_score": 4.2, "tier": "Gold" },
      "submitted_at": "2026-02-16T12:05:00Z"
    },
    ...
  ]
}

Webhooks (recommended for production)

Include a webhook_url in your query creation request, or set an account-level default with PUT /requestor/webhook. We POST a signed JSON payload on each response and on completion. See the Webhooks guide.

Rate Workers (Optional)

Rate each response 1–5 stars to improve the quality of future workers assigned to your queries. Workers with consistently low ratings are automatically suspended.

curl -X POST https://www.getabrain.ai/api/v1/requestor/queries/query_abc123/responses/resp_001/rate \
  -H "X-API-Key: gab_k_xxxxx" \
  -H "X-API-Secret: gab_s_xxxxx" \
  -H "Content-Type: application/json" \
  -d '{ "score": 5, "feedback_text": "Excellent, detailed feedback." }'

For AI Agents

GetABrain ships an MCP server (@getabrain/mcp-server) so agents in Claude Desktop, Cursor, Claude Code, or any MCP-compatible client can call ask_humanand friends directly as tools — no custom integration code required. Drop your API key/secret from Step 1 into the config below:

{
  "mcpServers": {
    "getabrain": {
      "command": "npx",
      "args": ["-y", "@getabrain/mcp-server"],
      "env": {
        "GETABRAIN_API_KEY": "gab_k_xxxxx",
        "GETABRAIN_API_SECRET": "gab_s_xxxxx"
      }
    }
  }
}

Building your own agent framework instead? Point any OpenAPI-aware tool generator (LangChain, OpenAI function calling, custom orchestrators) at the canonical spec and it will auto-generate typed tools for every endpoint:

https://www.getabrain.ai/api/v1/openapi.json

Prefer to browse it interactively first? See the API Reference.

All Query Types at a Glance

Choose the right type for your task. See the full Query Types reference for complete schemas.

Type	Use When You Need	Example
`text`	Open-ended written answers	"What do you think of this design?"
`multiple_choice`	Picking from options	"Which feature matters most?"
`rating_scale`	Numeric ratings (with optional media)	"Rate this product photo 1-10"
`image_comparison`	Choosing between images	"Which logo is more professional?"
`ranking`	Ordering items by preference	"Rank these features by importance"
`yes_no`	Binary decisions	"Is this headline effective?"
`sentiment`	Emotional analysis	"What is the sentiment of this review?"
`image_selection`	Selecting matching images	"Pick all photos suitable for a listing"
`free_form_text`	Long-form writing	"Write a product description"
`video_review`	YouTube video feedback	"Review this tutorial video"
`audio_review`	Audio content feedback	"Review this podcast episode"
`image_analysis`	Detailed image analysis	"Describe what you see in this photo"
`ab_test`	Comparing two variants	"Which landing page converts better?"
`voice_capture`	Worker-recorded voice note	"Record a 30-second reaction to this ad"
`video_capture`	Worker-recorded video	"Film yourself unboxing this product"
`photo_capture`	Worker-captured photo	"Take a photo of your desk setup"

More Examples

Sentiment Analysis

{
  "type": "sentiment",
  "title": "Analyze tweet sentiment",
  "content_data": {
    "question": "What is the sentiment of this text?",
    "subject": "Just tried the new feature and it completely changed my workflow!",
    "require_reasoning": true
  },
  "required_responses": 5,
  "bid_amount_cents": 10
}

A/B Test with Images

{
  "type": "ab_test",
  "title": "Landing page A/B test",
  "content_data": {
    "question": "Which design do you prefer and why?",
    "variant_a": {
      "description": "Minimalist hero with large image",
      "image_url": "https://example.com/design-a.png"
    },
    "variant_b": {
      "description": "Feature-rich layout with testimonials",
      "image_url": "https://example.com/design-b.png"
    }
  },
  "required_responses": 20,
  "bid_amount_cents": 25
}

Rating Scale with Media

{
  "type": "rating_scale",
  "title": "Rate this product video",
  "content_data": {
    "question": "How would you rate the overall quality?",
    "scale_type": "stars",
    "scale_min": 1,
    "scale_max": 5,
    "scale_labels": { "min_label": "Poor", "max_label": "Excellent" },
    "criteria": ["Visual quality", "Audio clarity", "Content value"],
    "media_video_url": "https://www.youtube.com/watch?v=dQw4w9WgXcQ"
  },
  "required_responses": 10,
  "bid_amount_cents": 30
}

Image Comparison

{
  "type": "image_comparison",
  "title": "Logo design preference",
  "content_data": {
    "question": "Which logo is more professional?",
    "comparison_type": "preference",
    "require_reasoning": true,
    "images": [
      { "id": "a", "url": "https://example.com/logo-a.png", "label": "Option A" },
      { "id": "b", "url": "https://example.com/logo-b.png", "label": "Option B" }
    ]
  },
  "required_responses": 10,
  "bid_amount_cents": 20
}

Common Query Parameters

Field	Type	Required	Description
`type`	string	Yes	One of the 16 query types
`title`	string	Yes	Short title shown to workers
`description`	string	No	Longer description for workers
`content_data`	object	Yes	Type-specific data (question, images, etc.)
`required_responses`	integer	Yes	How many worker responses you want
`bid_amount_cents`	integer	Yes	Pay per response in cents (min 5)
`bonus_amount_cents`	integer	No	Extra pay for thoughtful comments
`expires_in_hours`	integer	No	Auto-expire after N hours
`min_worker_quality`	number	No	Min quality score 0-5 (0 = any)
`webhook_url`	string	No	Public https URL to receive signed event notifications