Skip to main content
The integration has two components:
  • MCP server: Authenticates with the CatchAll API and exposes tools in Claude.
  • SKILL file: Teaches Claude when and how to use those tools, including query structure, validators, enrichments, polling, and result formatting.

​
Before you start

​
Connect MCP server

1

Open connectors

Go to claude.ai/customize/connectors. Click + and select Add custom connector.
2

Configure connection

Fill in the Add custom connector dialog:
  • Name: CatchAll
  • Remote MCP server URL:
https://catchall-mc.fastmcp.app/mcp?apiKey=YOUR_CATCHALL_API_KEY
3

Add and verify

Click Add. Verify that CatchAll appears under Web in your connectors list.
4

Test connection

Open a new chat and type a CatchAll query, for example: “Find AI company acquisitions in the last 7 days, limit 5”. Claude should call the CatchAll tools and return structured results.
The MCP URL contains your API key. Treat it as a secret and do not share it or commit it to version control.

​
Add SKILL

The SKILL file teaches Claude how to use CatchAll effectively. It includes the main prompt in SKILL.md along with validator and enrichment references, monitor scheduling guidance, and the OpenAPI specification.
1

Download SKILL archive

Download CatchAll-SKILL.zip from the integrations repository.
2

Upload to Claude

Go to claude.ai/customize/skills. Click +, select Upload a skill, and select CatchAll-SKILL.zip.Claude installs the SKILL and applies it to all your conversations and projects.

​
Available tools

The MCP server exposes tools for jobs (initialize, submit, poll, pull results, continue) and monitors (create, update, list, enable, disable, pull results). For the full tool reference, see MCP server > Available tools.

​
Example prompts

With the MCP connector and SKILL installed, describe a research task in chat. Claude submits jobs, tracks progress, and returns structured results.
Find Series B fintech funding rounds announced in the last 14 days, limit 20.
Search for AI chip export restrictions in the last 30 days and summarize the key regulatory actions.
Find pharma acquisitions over $1B this month, then create a table sorted by deal value.
Look for executive leadership changes at Fortune 500 companies last week and identify any patterns.
Specify a date range and a record limit in your prompt to scope results. For example: “in the last 14 days, limit 20”.

​
Python agent

To use CatchAll programmatically without the MCP connector, use the Anthropic SDK to execute CatchAll API calls directly in an agentic loop.

​
Setup

pip install -r requirements.txt
Set environment variables: 
export CATCHALL_API_KEY=YOUR_CATCHALL_API_KEY
export ANTHROPIC_API_KEY=YOUR_ANTHROPIC_API_KEY

​
Basic agent example

The agent uses hardcoded tool definitions and a standard agentic loop: submit, wait, pull with polling. 
client = anthropic.Anthropic(api_key=os.environ["ANTHROPIC_API_KEY"])

response = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=4096,
    tools=TOOLS,  # Hardcoded tool definitions in claude_agent_example.py
    messages=[
        {"role": "user", "content": "Find pharma M&A deals in the last 7 days, limit 20"}
    ]
)
For the complete implementation, including polling and result handling, see the agent example on GitHub.

​
SKILL-native agent example

The SKILL-native agent loads SKILL.md as the system prompt and generates tool definitions automatically by parsing its content. When the SKILL changes, the agent picks up those changes without code updates.
This example uses keyword matching to detect parameters in SKILL.md. Test after any SKILL update to verify that all parameters are detected correctly.

​
See also

MCP server

Server architecture, authentication, and tool definitions

Write effective queries

Get better results from CatchAll jobs

API reference

Full endpoint documentation

SKILL source

SKILL.md and reference files