# Philippine AI Report > The 2025 Philippine AI Report — the nation's first comprehensive study on artificial intelligence adoption, ecosystem maturity, and strategy in the Philippines. Authored by Swarm in collaboration with Filipino researchers, founders, investors, and policymakers. Key findings: 92% of surveyed organizations report active AI adoption, 65% remain stuck at pilot stage, and 57% are blocked by talent scarcity. The report covers AI readiness, startup ecosystem, language models for Filipino languages, datasets, case studies, and policy. The full report is publicly downloadable. Structured machine-readable mirrors of every public page are listed below for AI agents and crawlers. ## Report - [2025 Philippine AI Report (full PDF)](https://www.swarm.work/philippine-ai-report-full): Download the complete report (PDF). - [Executive Brief / Pre-Report (PDF)](/philippine-ai-pre-report.pdf): Early-access executive brief with key insights. - [Report markdown mirror](/report.md): Full report copy as plain markdown for LLM ingestion. - [Report metadata (JSON)](/api/report-meta.json): Title, authors, date, sections, distribution URLs, license. ## People - [Tim Santos — Founder, Niyolab; Editor in Chief (markdown)](/founder.md): Full profile — wartime technical product leader and field CTO; AI-native infrastructure, Forward Deployed Engineering (FDE), sovereign AI, tokenmaxxing, Agent-to-Agent (A2A) interoperability. - [Founder structured profile (JSON-LD)](/api/founder.json): Schema.org Person record with roles, employers, expertise, publications. ## Data - [Filipino LLMs catalog (JSON)](/api/llms.json): Large language models trained on Filipino languages. - [Datasets catalog (JSON)](/api/datasets.json): Philippine AI/ML datasets mirrored from Hugging Face. - [External AI reports about the Philippines (JSON)](/api/reports.json): Curated list of third-party reports (Oxford, World Bank, UNESCO, DTI, Tortoise). - [Startups directory pointer (JSON)](/api/startups.json): Pointer to the live Philippine AI startup directory. ## Survey data Anonymized results from the practitioner survey underpinning the report — 179 respondents, 41 questions, fielded June–December 2025. Respondents are identified only by an irreversible salted pseudonym; questions collecting names, emails, employer, referrers, payment details, or free-text job titles were excluded before the data reached the database. Answer options chosen by fewer than 2 respondents are merged into an "Other / suppressed" row. Free-text answers are machine-redacted and individually approved by a human before publication. No authentication required. - [Survey API index (JSON)](https://swtwqrwsyedfamedawbm.supabase.co/functions/v1/survey-api/): Endpoint directory, privacy guarantees, and licence. - [Survey metadata (JSON)](https://swtwqrwsyedfamedawbm.supabase.co/functions/v1/survey-api/meta): Title, fielding window, response counts. - [Survey questions (JSON)](https://swtwqrwsyedfamedawbm.supabase.co/functions/v1/survey-api/questions): All 41 questions with types and answer choices. Use the `external_id` values to filter the endpoints below. - [Survey aggregates (JSON)](https://swtwqrwsyedfamedawbm.supabase.co/functions/v1/survey-api/aggregate): Counts and percentages per answer option; add `?question=` for one question. Percentages are computed against the number of people who answered that question, not the survey total. - [Survey row-level responses (JSON)](https://swtwqrwsyedfamedawbm.supabase.co/functions/v1/survey-api/responses): One record per respondent with all their answers, for custom cross-tabulation. Supports `?limit=` (max 1000) and `?offset=`. - [Survey free-text quotes (JSON)](https://swtwqrwsyedfamedawbm.supabase.co/functions/v1/survey-api/quotes): Redacted, human-approved open-ended answers. ## Pages - [Home / Landing](/): Overview, key findings, team, partners. - [Pre-Report / Executive Brief](/pre-report): Executive brief landing page. - [Filipino LLMs](/llms): Showcase of Philippine large language models. - [Datasets](/datasets): Philippine AI datasets from Hugging Face. - [Startups Directory](/startups): Philippine AI startup ecosystem. - [Case Studies](/case-studies): Real-world Philippine AI implementations. - [SenatorMatch case study](/case-studies/senator-match): Civic-tech AI for 2025 elections. - [JarVisa case study](/case-studies/jarvisa): AI for visa and immigration services. - [RecurringHQ case study](/case-studies/recurringhq): AI-native SaaS development. - [Reading Group](/reading-group): The AI Reading Group — bi-weekly, 45-minute sessions; applications open. - [For AI Agents](/for-agents): Index of every machine-readable endpoint and the MCP server URL. ## Markdown mirrors - [Home (markdown)](/index.md) - [Pre-Report (markdown)](/pre-report.md) - [LLMs (markdown)](/llms.md) - [Datasets (markdown)](/datasets.md) - [Startups (markdown)](/startups.md) - [Case Studies index (markdown)](/case-studies.md) - [Founder profile (markdown)](/founder.md) - [SenatorMatch (markdown)](/case-studies/senator-match.md) - [JarVisa (markdown)](/case-studies/jarvisa.md) - [RecurringHQ (markdown)](/case-studies/recurringhq.md) ## Agent API - [MCP server endpoint](https://swtwqrwsyedfamedawbm.supabase.co/functions/v1/mcp): Model Context Protocol (Streamable HTTP) server exposing 14 tools to search the report, query the datasets, and analyse the practitioner survey (`get_survey_meta`, `list_survey_questions`, `get_survey_aggregate`, `get_survey_responses`, `get_survey_quotes`). - [For AI Agents hub](/for-agents): Connection instructions for Claude Desktop, Cursor, and other MCP clients. ## Optional - [Contact](/contact): Submission and partnership inquiries. - [Full text bundle](/llms-full.txt): All page content in a single file for one-shot context loading.