5
Commitments we publish
100%
Coverage of AI operations
0
Fabrications caught
10
Dimensions per match
Where we are this quarter.
The Transparency Report publishes one quarterly. This section summarises the headline operating numbers; the next three tabs drill into Financials, AI Usage, and Verification Stats.
Publishes Q2 2026
Organizations with claimed profiles
Running count, not cumulative.
Publishes Q2 2026
Donors on the platform
All personas combined.
Publishes Q2 2026
Grant recommendation packages prepared
Catylst prepares; your DAF custodian executes.
Quarterly
Reporting cadence
Next report: Q2 2026 — July 15, 2026.
Last published: April 15, 2026. Next report: Q2 2026 — July 15, 2026.
High-level operating disclosures.
Phase A operating disclosures at the band level. Specific figures publish in each quarterly report; audited financials begin Year 2.
Revenue
Pre-revenue (Phase A launch)
Paid tiers activate post-launch; see /pricing for the tier rack.
Headcount
4–6 team members
Founder + engineering + product + design + advisor network.
Funding status
Bootstrapped / pre-seed
No institutional lead investor as of the report date.
Runway posture
Conservative
Catylst does not custody donor funds and does not monetize faith data.
Commitment: Catylst publishes annual audited financials starting Year 2. Year-1 disclosures stay at the band level in these quarterly reports.
As of April 15, 2026.
What Catylst sends to which model — and what it keeps.
Representative operations, not exhaustive. The full list of AI calls publishes in each quarterly report with request-volume counts and retention details.
Match explanation generation
Produces the cited, ten-dimension explanation for a donor-organization pair.
Data sent:Donor profile fields + organization profile fields that the donor has consented to match on.
Retention:Zero retention at the model provider (ephemeral request).
Deal-room diligence summary
Summarises uploaded diligence documents inside a deal room for user review.
Data sent:Documents the user uploaded to the deal room; metadata only where possible.
Retention:Zero retention at the model provider; caching limited to the user's session.
Organization categorization
Classifies a claimed organization against cause, geography, and program taxonomies.
Data sent:Public-facing profile fields the organization has published.
Retention:Zero retention at the model provider.
Question answering on organization profiles
Answers donor questions by citing published organization profile content.
Data sent:The donor's question plus the cited org-profile excerpts the answer draws on.
Retention:Zero retention at the model provider.
Trust-badge eligibility review
Advises badge reviewers whether an organization has met documented criteria.
Data sent:Structured verification artifacts submitted by the organization.
Retention:Zero retention at the model provider; review outcomes logged in Catylst audit trail.
Profile embeddings
Produces vector embeddings that power the semantic discovery feed.
Data sent:Public profile text fields only.
Retention:Vectors persist in Catylst's vector store; provider-side retention is zero.
Cross-model validation
Runs a second-model sanity check on high-stakes recommendations before emission.
Data sent:The primary-model output plus the cited profile fields it referenced.
Retention:Zero retention at the model provider.
Faith-content sensitivity review
Flags pastoral sensitivity concerns for human advisor review prior to publication.
Data sent:Draft content plus relevant pastoral-tradition metadata.
Retention:Zero retention at the model provider; flagged items logged for advisor review.
Claude 4.6 (Anthropic)
Primary generation model for explanations, summaries, and Q&A.
GPT-5.2-pro (OpenAI)
Cross-check model for high-stakes recommendations.
Voyage-3 (Voyage AI)
Embedding model for the semantic discovery index.
Monthly AI spend
Under $2,500 / month (Phase A)
Disclosed at the band level during Phase A; precise spend figures publish in each quarterly report.
As of April 15, 2026.
How organizations have earned trust badges.
The seven-badge trust architecture (identity, nonprofit status, diligence, enhanced reporting, financial docs, network endorsement, advisor vetting). Per-badge verification rates publish in each quarterly report.
- Identity Verified
- Nonprofit Status
- Diligence Ready
- Enhanced Reporting
- Financial Docs Reviewed
- Network Endorsed
- Advisor Vetted
Identity Verified
Publishes Q2 2026
% of organizations with identity verification completed.
Nonprofit Status
Publishes Q2 2026
% with active IRS 501(c)(3) status cross-referenced to public records.
Financial Docs Reviewed
Publishes Q2 2026
% with audited financials reviewed and matched to the public 990.
Diligence Ready
Publishes Q2 2026
% with the full diligence artifact set on file.
Enhanced Reporting
Publishes Q2 2026
% publishing outcome metrics on a committed public cadence.
Network Endorsed
Publishes Q2 2026
% endorsed by a denominational or funding network.
Advisor Vetted
Publishes Q2 2026
% recommended by an independent advisor on the platform.
Breakdown by organization type: The first quarterly report also splits verification rates by organization type (faith-led nonprofit, secular nonprofit, church, school, foundation grantee). No paid placement and no badge is ever sold — every badge is earned.
As of April 15, 2026.