Grant matching and workflow support
through a dedicated MCP lane.

What it is

The grant finder is a private MCP workflow for pre-submission grant work. It is separate from the published pack MCP because it runs a guided workflow, not a generic data query surface.

Start with a plain-language project description. The MCP builds a structured session, asks for missing facts when they matter, and then moves through prospect research, target selection, memo support, budget framing, and submission checks.

Endpoint

https://app.daedalmap.com/mcp-private/grants

Bearer token required.

What it does

• Match a project against plausible grant programs
• Separate direct fits from blocked-but-workable paths
• Summarize precedent and current opportunity posture where supported
• Help move a shortlist into memo, budget, and submission-support work

What it does not do

• Post-award compliance
• Reporting, reimbursement, or closeout work
• Win-probability claims

First flow

1. Connect your MCP client to https://app.daedalmap.com/mcp-private/grants.
2. Call grant_intake_or_update_project with project_input.
3. Use the returned session_id in grant_analyze_current_phase.
4. If phase 1 returns a shortlist, call grant_select_target_program.
5. Call grant_analyze_current_phase again for memo, budget, or submission-support work.

First call

Start with grant_intake_or_update_project. Put the user's project description into project_input. If you already know structured facts, pass them in profile_overrides.

{
  "name": "grant_intake_or_update_project",
  "arguments": {
    "project_input": "We are a nonprofit 501c3 working on wildfire debris removal, wood products reuse, and disaster recovery in Appalachia.",
    "profile_overrides": {
      "recipient_loc_id": "USA-NC",
      "place_of_performance_loc_id": "USA-NC",
      "has_government_partner": true
    }
  }
}

Second call

Use the returned session_id in grant_analyze_current_phase.

{
  "name": "grant_analyze_current_phase",
  "arguments": {
    "session_id": "gs_example"
  }
}

What good input looks like

The strongest first input says who the applicant is, where the work will happen, what kind of work is being funded, and whether there is already a government partner or match funding.

• Applicant type: nonprofit, higher ed, tribal government, local government, and similar
• Geography: where the applicant is based and where the work will happen
• Project type: disaster recovery, forestry, research, infrastructure, resilience, and similar
• Known constraints: partner requirements, match funding, target award amount, project length

Clarification is normal

The grant finder asks clarification questions when key facts are missing. That is part of the contract, not an error path.

The most important missing-field split is often recipient_loc_id versus place_of_performance_loc_id. The applying organization and the funded work are not always in the same place.

Phase 2 and phase 3 follow-up

After you select a target program, ask for the next task directly.

• Memo support: Build memo talking points for this target.
• Budget framing: Help me build the budget framework.
• Submission check: Is this program open right now?