kindora.co

Kindora MCP Server - Access foundation and funder data. This server provides tools to search and analyze philanthropic foundations using IRS 990 data. Available tools: FREE TIER (No authentication required): - search_funders: Look up grantmaking organizations by name, topic, or location - search_open_grants: Find open grant opportunities by TOPIC, cause area, or keywords (full-text search across program names, descriptions, focus areas, beneficiary types, and geographic focus; supports natural language queries, quoted phrases, and -term exclusion; use state param for geographic targeting) - search_funder_jobs: Find OPEN PHILANTHROPY JOBS at foundations — program officers, grants managers, foundation CEOs, etc. Filter by category (grantmaking, program_leadership, executive_leadership, etc.), state, country (funder HQ — "US"/"GB"/"Germany"; the corpus spans European funders), remote status, or funder_ein. Backed by a weekly scrape + GPT-5.4-mini classification. - get_funder_profile: Get detailed profile for a specific foundation - get_990_summary: Get IRS 990 filing financial summary - get_foundation_grants: View grants made by a foundation - get_ntee_codes: Browse NTEE classification codes - get_funder_stats: Get aggregate giving statistics CRITICAL — CHOOSING THE RIGHT SEARCH TOOL: - User wants ACTIVE grant opportunities or RFPs for a topic/cause area (e.g., "criminal justice reform", "youth education", "climate resilience") → use search_open_grants - User wants FUNDER ORGANIZATIONS aligned to a topic/cause area, including funders that may not have a live RFP right now → use search_funders - User wants JOBS in philanthropy (program officers, grants managers, foundation CEOs, etc.) → use search_funder_jobs - User names a SPECIFIC ORGANIZATION (e.g., "Ford Foundation", "MacArthur") → use search_funders - User wants to BROWSE funders by location or size → use search_funders - When exploring a cause area, it is often helpful to run BOTH: search_funders for aligned funders and search_open_grants for active opportunities. BEST PRACTICES for search_open_grants: - Always use the state parameter when the user has a known US location - For non-US geographies (e.g. India, Kenya, Mexico, Global), use the country parameter — pass the country name as plain text, e.g. country="India". Mixing state with a non-US country is rejected because state implies US-only filtering. - Use natural language queries that describe what you're looking for (e.g., "youth after school programs", "affordable housing for seniors") - Quoted phrases work for exact matching: '"mental health" youth' - Use -term to exclude: "education -higher" excludes higher education results - Run multiple targeted searches with different keywords rather than one broad search - Combine state + query (or country + query) for the most relevant results - Omit query entirely to browse open programs ordered by upcoming deadline - Do NOT filter by source="government" unless specifically asked — the foundation corpus is much larger and most topic searches return better results without filtering. Omit the source parameter to search both sources together. BEST PRACTICES for search_funders: - Use the `state` parameter to filter by the funder's US HQ state. - Use `grantee_country_codes` (FIPS 10-4 codes — "UK", "IN", "KE", "SF", etc.) to restrict to funders whose actual grantees are based in a given non-US country. This is the right filter when the user asks "who funds work in India?" or "find me UK-grantee funders" — it joins against the grantee-distribution materialized view rather than the funder's own HQ. Pass a list of codes; multiples OR together. Cannot be combined with ILIKE-only single-word name lookups (e.g. "Gates"); the filter forces the hybrid path automatically when set. All tools are read-only and rate-limited. Exact request ceilings vary by traffic tier and tool cost; anonymous users should expect roughly 100 requests/hour overall with tighter caps on expensive browse-heavy paths. FAIR USE & BATCH INTEGRATIONS: Interactive discovery (a person or assistant researching funders) is free and welcome. Automated, scheduled, or batch integrations are welcome too, but MUST self-identify by sending an 'X-Kindora-Client: <your-app-name>' header on every request. Identifying yourself pays: on serverless platforms that rotate egress IPs (e.g. Supabase Edge Functions), apps that do not identify themselves share one small pooled allowance, while identified apps draw from the platform's larger shared allowance. The label is attribution, not authentication. DEDICATED ACCESS: integrators who need more than the shared free allowances can register a Kindora account at https://www.kindora.co/developers and send a service key ('Authorization: Bearer kdra_org_...') for a dedicated org lane with substantially higher velocity limits. For sustained batch volume or data licensing, email support@kindora.co — we are friendly to integrators and would rather talk than throttle.