MCP Integration

​

Quick Install

codex mcp add tinyfish --url https://agent.tinyfish.ai/mcp

grok plugin install tinyfish --trust

grok /mcps

npx -y install-mcp@latest https://agent.tinyfish.ai/mcp --client claude-code

opencode mcp add tinyfish --url https://agent.tinyfish.ai/mcp

npx -y @tiny-fish/cli@latest connect command-code

https://agent.tinyfish.ai/mcp/chatgpt

npx -y install-mcp@latest https://agent.tinyfish.ai/mcp --client cursor

npx -y install-mcp@latest https://agent.tinyfish.ai/mcp --client windsurf

codex mcp add tinyfish --url https://agent.tinyfish.ai/mcp

grok mcp add --transport http tinyfish https://agent.tinyfish.ai/mcp

[mcp_servers.tinyfish]
url = "https://agent.tinyfish.ai/mcp"

claude mcp add --transport http tinyfish https://agent.tinyfish.ai/mcp

opencode mcp add tinyfish --url https://agent.tinyfish.ai/mcp

{
  "mcp": {
    "tinyfish": {
      "type": "remote",
      "url": "https://agent.tinyfish.ai/mcp"
    }
  }
}

npx -y @tiny-fish/cli@latest connect command-code --api-key sk-tinyfish-...

{
  "mcpServers": {
    "tinyfish": {
      "transport": "http",
      "url": "https://agent.tinyfish.ai/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

https://agent.tinyfish.ai/mcp/chatgpt

{
  "mcpServers": {
    "tinyfish": {
      "url": "https://agent.tinyfish.ai/mcp"
    }
  }
}

{
  "mcpServers": {
    "tinyfish": {
      "serverUrl": "https://agent.tinyfish.ai/mcp"
    }
  }
}

​

Available Tools

​

Web Automation

• run_web_automation — Execute multi-step web automation given a URL and natural language goal. Streams progress in real-time. Use this for tasks that require clicking, navigating, filling forms, logging in, or interacting with a website. For authenticated recurring workflows, pass use_profile: true; add profile_id to select a specific Browser Context Profile.
• run_web_automation_async — Same as run_web_automation but returns a run_id immediately without waiting. Use for long-running tasks where you don’t need real-time progress. Poll with get_run. Supports the same use_profile, profile_id, and use_vault controls.
• get_run — Retrieve details of a specific automation run by ID. Returns status, result, and error info.
• list_runs — List your automation runs with optional filtering by status and pagination.
• cancel_run — Cancel a running or pending automation run by ID. Idempotent — already-terminal runs return their current status.

​

Batch Automation

• batch_status — Check the status of multiple runs at once (up to 8). Poll every 30–60 seconds until all_terminal is true.
• batch_cancel — Cancel multiple running or pending runs by their IDs (up to 8). Idempotent.

​

Web Search

• search — Search the web and get structured results with titles, snippets, and URLs. Use this to find information or discover URLs before automating a specific site. Optional filters: recency_minutes, after_date, before_date, domain_type ("web" | "news" | "research_paper"), pub_year_min / pub_year_max. When domain_type=research_paper, recency_minutes, after_date, and before_date are not supported — use pub_year_min / pub_year_max instead to scope by publication year (inclusive integers, 0-9999, with pub_year_min <= pub_year_max when both are set). You can also narrow or bias results with location (country code, e.g. "US"), language (language code, e.g. "en"), and include_domains / exclude_domains (comma-separated domain allow/deny lists, combined length capped). Use page (0-indexed, 0-10) to paginate. You can also pass an optional purpose — a short statement of why you’re searching — to give the search additional intent signal for better-quality results.
• get_search_usage — List past search usage records with filtering by date range and status.

​

Content Extraction

• fetch_content — Fetch web pages, render JavaScript-heavy pages when needed, and extract clean, structured content. Preferred over run_web_automation when you only need to read page content. Supports up to 10 URLs per request. Control the output with format ("markdown" default | "html" | "json"), links and image_links (extract outbound/image URLs), ttl (cache freshness tolerance in seconds), and per_url_timeout_ms (per-URL wall-clock budget). Scope extraction with include_selectors / exclude_selectors (1–20 CSS selectors each). For conditional requests, pass if_none_match / if_modified_since (single URL only) and set include_etag_and_last_modified to get the validators back. You can also pass an optional purpose — a short statement of why you’re fetching the URLs — to give the fetch additional intent signal for better-quality results.
• list_fetch_usage — List past fetch requests with filtering by date range and status. Returns metadata only.

​

Deep Research

• start_deep_research — Start a deep research run that searches, reads, and synthesizes many sources into a cited report. Pass query and an optional mode: "deep" (default) or "max" (deeper, with a larger step budget, for when you ask for maximum depth). Returns a research_run_id and a report_url immediately; sign in to open the report there any time.
• get_deep_research — Check a run by research_run_id. Waits up to about 25 seconds, then returns its status, progress and report_url. When output_ready is true, the result includes the report and its citations.
• list_deep_research — List your runs, newest first, with status and report_url; filter by status, text or date and page with cursor, like GET /v1/research-run.
• cancel_deep_research — Stop a run by research_run_id. Idempotent — finished runs return their current status.

​

Monitors

• create_monitor — Create a recurring Monitor and get its baseline result immediately. Set type to "search" with config.query to watch a topic, or "fetch" with config.url to watch a page. schedule_cron is a 5-field cron, UTC unless prefixed with CRON_TZ=<zone>. Optional: name, webhook_url, purpose. Scheduled runs post to webhook_url; Page payloads say whether the page changed and include the diff. When the page changed and purpose is set, they also judge the change against it (payload).
• list_monitors — List your Monitors with their id, type, target, schedule, and status.
• get_monitor — Get one Monitor by monitor_id.
• pause_monitor — Pause a Monitor’s schedule by monitor_id.
• resume_monitor — Resume a paused Monitor’s schedule by monitor_id.
• run_monitor — Run a Monitor once now by monitor_id and return that result.
• cancel_monitor — Delete a Monitor and its schedule by monitor_id.

​

Datasets

• create_dataset — Build a dataset: a table of many entities with the same fields, each row found and filled from the web. Pass prompt, or a spec returned by plan_dataset once you have reviewed the columns, and an optional target_rows (1-250, default 100). Returns dataset_id and run_id immediately.
• plan_dataset — Design the columns for a prompt without creating anything. Edit the returned spec and pass it to create_dataset.
• get_dataset — Check a dataset by dataset_id. Waits up to about 25 seconds, then returns its status, row_count, and the first 50 rows. When output_ready is true the rows stand; call get_dataset_rows for more.
• get_dataset_rows — One page of rows, up to 500, in column order. Pass the previous next_cursor as after for the next page.
• list_datasets — Your datasets, newest first, with status and row counts.
• populate_dataset — Fill an existing dataset again by dataset_id: raise target_rows to add rows, or set reset to rebuild it. Refused while a run is live.
• cancel_dataset_run — Stop a run by dataset_id and run_id. Rows already added stay.
• delete_dataset — Delete a dataset and its rows by dataset_id. Cancel any live run first.

​

Browser Sessions

• create_browser_session — Create a remote, stealth Chrome browser session and get CDP connection details. Use for direct programmatic browser control with Playwright, Puppeteer, or Selenium.
• list_browser_sessions — List your browser sessions with optional filtering by session ID, time range, and status.
• close_browser_session — Close a browser session by ID to release it early instead of waiting for its inactivity timeout. Idempotent — already-ended sessions return success.

​

Wallet

• get_wallet — Read-only. Returns your current wallet balance, auto-reload state, per-product contract rates, and any in-flight top-up. Auto-reload changes happen in the dashboard.
• top_up_wallet — Adds money to your wallet, paid by the agent over the Machine Payments Protocol: the first call returns a payment challenge, the second call presents the paid credential. Out-of-funds refusals point at it, and your wallet asks you to approve the payment. Accounts outside the rollout don’t see this tool in the tool list, and calling it anyway returns Unknown tool: top_up_wallet. See Agent Payments.

​

Profile-Aware Automation

{
  "url": "https://app.example.com/dashboard",
  "goal": "Summarize the account health dashboard",
  "use_profile": true,
  "use_vault": true
}

{
  "url": "https://app.example.com/dashboard",
  "goal": "Summarize the latest invoice, including the amount due and payment deadline.",
  "use_profile": true,
  "profile_id": "prof_abc123def4567890"
}

​

Example Usage

"Go to https://example.com and extract the product prices"

"Use my saved Salesforce profile to summarize the dashboard"

"Fill out the vendor onboarding form using these company details, then stop before submitting."

"Find overdue invoices in my billing dashboard and list each customer, amount due, and due date."

"Search for the best Italian restaurants in San Francisco"

"Scrape the pricing pages from these 5 competitor websites at the same time"

"Fetch https://docs.stripe.com/api/charges and summarize the API parameters"

"Create a browser session on https://example.com so I can control it with Playwright"

"Do deep research on long-duration energy storage approaches and give me a cited report"

"Monitor https://example.com/pricing every morning and send each result to https://hooks.example.com/pricing"

"Build a dataset of specialty coffee shops in Oakland with their website and rating"

​

Rates

​

Authentication

​

Before You Start

1. claude.ai — your Claude/Anthropic account
2. agent.tinyfish.ai — your TinyFish account

​

Authentication Flow

​

Troubleshooting

​

Related