Clay Navigator charges $0.36 per run. Huntr Deep charges $0.046. 7.8× cheaper. No subscription. Pay for returned data. Use from Clay, n8n, or any HTTP client.
// 15 endpoints · one api key
AI research agent plus direct scraping endpoints. One key, one pricing model, no subscription. Rate limit: 5 requests/second per API key.
// ai agent
// POST /research · the ai research agent
Send a plain-text prompt. The agent plans which endpoints to call, executes them in parallel, and synthesizes a result. The same prompt — two completely different approaches.
clayClay generates a 500-word prompt with 7 fallback strategies because it has no LinkedIn posts API:
"If the CEO's recent posts are not publicly visible without login, use credible secondary sources (e.g., reputable media coverage or the CEO's cross-posts on other public platforms like X) to infer the content of recent posts."
google_search → "CEO of LinkedIn 2026" → Ryan Roslanskyperson_url_finder → linkedin.com/in/ryanroslanskylinkedin_profile_posts → 10 real posts, real text, real engagement// try it
Click any prompt. Watch Huntr plan, execute, and synthesize — in real time.
// huntr vs claygent
One table. Every number real.
clay// pricing
Top up credits, use them when you need. No monthly minimums.
Agent Pricing
AI plans, calls capabilities, and synthesizes a result. One prompt in, one answer out.
Search + scrape — the full research toolkit
Company enrichment + advanced extraction
Full people enrichment pipeline
| Model | Lite | Standard | Deep | Clay Claygent | vs Clay (Huntr Deep) |
|---|---|---|---|---|---|
| gpt-5-nano | $0.015 | $0.021 | $0.037 | $0.06 | 1.6× |
| gemini-2.5-flash-lite | $0.015 | $0.022 | $0.037 | — | N/A |
| deepseek-v3.2 | $0.015 | $0.022 | $0.038 | — | N/A |
| gpt-4.1-nano | $0.015 | $0.022 | $0.037 | $0.06 | 1.6× |
| gpt-4o-mini | $0.015 | $0.022 | $0.039 | $0.06 | 1.5× |
| gpt-5-mini | $0.017 | $0.026 | $0.044 | $0.06 | 1.4× |
| gpt-4.1-mini | $0.017 | $0.026 | $0.045 | $0.06 | 1.3× |
| gemini-2.5-flash All tiers default | $0.018 | $0.027 | $0.046 | $0.06 | 1.3× |
| gemini-3-flash | $0.019 | $0.030 | $0.050 | ~$0.06 | 1.2× |
| claude-haiku-4-5 | $0.023 | $0.038 | $0.062 | ~$0.32 | 5.2× |
| gemini-2.5-pro | $0.029 | $0.050 | $0.080 | ~$0.20 | 2.5× |
| gpt-5 | $0.029 | $0.050 | $0.080 | ~$0.06 | Clay cheaper |
| gpt-4.1 | $0.030 | $0.052 | $0.083 | ~$0.41 | 4.9× |
| gpt-4o | $0.034 | $0.060 | $0.095 | ~$0.12 | 1.3× |
| gemini-3.1-pro | $0.034 | $0.060 | $0.095 | ~$0.32 | 3.4× |
| gpt-5.4 | $0.039 | $0.070 | $0.110 | ~$0.17 | 1.5× |
| claude-sonnet-4-6 | $0.041 | $0.074 | $0.116 | ~$0.25 | 2.2× |
| claude-opus-4.6 | $0.059 | $0.110 | $0.170 | ~$0.42 | 2.5× |
Price = tier base cost (infra + triage LLM + overhead) + synthesis model cost at token caps (Lite: 4K/1K · Standard: 8K/2K · Deep: 12K/3K) + $0.008 margin.
"vs Clay" ratio = Clay Claygent price ÷ Huntr Deep price. Clay Claygent costs are per Clay's published pricing. Clay prices marked ~ are estimates from their model tier documentation.
Direct API pricing
Standalone endpoints — no agent overhead. Search is per returned result; enrichment is charged only when data is found.
// list building
// person enrichment
// company enrichment
// page enrichment
Clay is a spreadsheet UI with AI enrichment built in. Huntr is an API — send a prompt or call an endpoint, get structured data back. No UI, no seats, no subscription. Plug it into Clay, n8n, Make, or your own code.
Agent runs use fixed Lite, Standard, and Deep prices. Direct search endpoints charge per returned result, and enrichment endpoints charge only when the requested data is found. Top up credits when you need them. New accounts get $5 free.
Yes. Huntr is a plain HTTP API. In Clay use the HTTP enrichment block. In n8n or Make use an HTTP node. Send a POST to /research with your prompt, get JSON back. For async workflows, pass a callback_url and we'll POST the result when done.
Individual steps can fail without crashing the whole request. The agent collects whatever it found and synthesis always runs — you always get a response. If the entire request fails before synthesis, you're not charged.
Yes. Point Claude Desktop or Cursor at POST /mcp with your API key and Huntr appears as a tool. Run research queries directly from your AI assistant.