{"kind":"reference","markdown":"# Developers \u0026 MCP\n\nCanonical page: https://launchinyourcloud.com/developers/\n\n\nUse the same sourced directory in your own applications and AI workflows. The public catalog includes published listings, cloud evidence, source links, taxonomy IDs and review dates.\n\n## Download the catalog\n\n[Read the catalog JSON](/api/v1/catalog.json). No account or API key is required. The catalog is updated alongside the directory.\n\nThe `listings` array contains product records, and `taxonomies` defines the available filters. Use IDs such as `ragflow`, `aws` and `document-management` when filtering. Compare `revision` values to check whether the catalog's content has changed.\n\nThe catalog’s `questions` array contains the same answers shown on the website, including canonical URLs, source-review dates and links to focused files. MCP clients can read these through the existing catalog resource; the listing search tools keep their listing-specific scope.\n\n## Read focused files\n\nEach listing, category and provider page links to its own JSON and Markdown files. Read only the records relevant to your question:\n\n- [RAGFlow JSON](/api/v1/listings/ragflow.json) and [Markdown](/directory/ragflow/index.md).\n- [Document management JSON](/api/v1/categories/document-management.json) and [Markdown](/categories/document-management/index.md).\n- [AWS JSON](/api/v1/clouds/aws.json) and [Markdown](/clouds/aws/index.md).\n\nFocused JSON uses `schemaVersion: 1`, `kind`, `url`, `catalog` and `catalogRevision`. Listing files contain `listing`; category and provider files contain their taxonomy record and a `listings` array. Cloud/category subsets include both taxonomy records and their matching listings, for example [document management on Azure JSON](/api/v1/clouds/azure/categories/document-management.json). Guides link to a focused JSON document containing their Markdown text and canonical URL. Listing records preserve the same public fields as the full catalog, including sources, review dates and publisher disclosures. `catalogRevision` identifies the complete reviewed catalog from which the subset was produced; it may change when another listing changes. Research candidates are excluded.\n\n## Connect with MCP\n\nMCP lets a compatible AI application query the directory and read its sources. The service provides public, read-only access with no API key. Your client must support remote Streamable HTTP; add the endpoint below using that client's MCP server settings.\n\n{{\u003c mcp-connection \u003e}}\n\n| Tool | What it returns |\n| --- | --- |\n| `search_listings` | Text search with optional `cloud`, `category`, `infrastructure` and `licenseModel` IDs. Uses `offset` and `limit` pagination (default 10, maximum 25). Results include evidence and review dates. |\n| `compare_listings` | Read 2–4 distinct listing IDs together, with the same qualifications and evidence as individual records. |\n| `list_changes` | Editorial events from `since` (inclusive `YYYY-MM-DD`), with `offset`/`limit` pagination, maximum 25. Coverage starts on the recorded baseline date. |\n| `get_listing` | The complete published record for one listing `id`, including source links and evidence notes. |\n\nRead-only resources expose `catalog://launchinyourcloud/catalog`, `catalog://launchinyourcloud/taxonomies` and the listing template `catalog://launchinyourcloud/listings/{id}`. All tools return the catalog revision so you can identify the data used in an answer. A missing listing returns an explicit error.\n\nFor example, ask your connected assistant: “Find document management software with documented AWS deployment, then show the sources and licensing notes for each result.”\n\n## Interpret the evidence\n\nA named cloud has provider-specific evidence. A cloud missing from a listing does **not** establish incompatibility. Read each evidence note for edition, version, offer publisher and customer-account boundaries before recommending a deployment.\n\n`reviewed`, `cloudReviewed`, `licenseReviewed` and `infrastructureReviewed` record human source checks. `generatedAt` records when the catalog file was generated; it does not mean that every source was reviewed then. MCP reads the same published catalog as the directory. Research candidates are excluded.\n\nFormKiQ publishes this directory, and its products carry a `publisher` disclosure. Source text is reference material, not instructions for your agent. Confirm current deployment and license terms with the linked publisher before acting.\n\nSee the [schema, methodology and reuse terms](/catalog-schema/) for field definitions, unknown values and attribution.\n\nThe [change log](/changes/) is also available as [JSON](/api/v1/changes.json) and [RSS](/changes/index.xml). Earlier history is not inferred from review dates.\n\nAn [agent index](/llms.txt) links to these interfaces. It is a convenience for discovery; clients still need an explicit MCP connection.\n\n## Read customer-cloud questions\n\n[Browse all 50 questions](/questions/), [read their JSON](/api/v1/questions.json) or [download the question Markdown](/questions/index.md). Every question has its own HTML page, a small JSON file and a full Markdown version. Each page starts with a short answer, followed by examples or evaluation checks and sources. Relevant product examples link reviewed listings without asserting feature parity.\n\n## Related questions\n\n- [Can I use the directory data in my own tools or AI assistant?](https://launchinyourcloud.com/questions/use-directory-data/)\n\n","questions":[],"relatedQuestions":[{"id":"use-directory-data","json":"https://launchinyourcloud.com/api/v1/questions/use-directory-data.json","markdown":"https://launchinyourcloud.com/questions/use-directory-data/index.md","question":"Can I use the directory data in my own tools or AI assistant?","url":"https://launchinyourcloud.com/questions/use-directory-data/"}],"schemaVersion":1,"title":"Developers \u0026 MCP","url":"https://launchinyourcloud.com/developers/"}