Flyder Developer API

Integrate your own application with Flyder using a single API key. Run workflows, drive chat tasks, and retrieve answers from your indexed knowledge — with first-class support for both Vectorless RAG (PageIndex) and traditional Vector RAG (Vector DB).

Get Your API Key

Every request to the Flyder API is authenticated with a Flyder API Key. Once you're signed in, generate one from Settings > Flyder API Key as shown below.

PageIndex: Vectorless RAG

PageIndex indexes your documents into a navigable structure instead of chunking and embedding them. Queries are answered by an LLM reasoning over that structure and pulling exact pages on demand — no vector database, no chunking artifacts, no embedding model to manage. It's the fastest-growing alternative to embedding-based RAG, and Flyder's API exposes it directly:

• POST /api/pageindex/documents — upload a PDF, Markdown, or text file for indexing
• GET /api/pageindex/jobs/{job_id} — poll indexing status
• GET /api/pageindex/documents — list your indexed documents
• GET /api/pageindex/documents/{doc_id}/structure — fetch a document's section/page outline
• GET /api/pageindex/documents/{doc_id}/pages?pages=5-7 — fetch raw page content
• POST /api/pageindex/query — ask a question and get a reasoned answer across one or more documents

Vector DB: Vector RAG

For workloads that want classic embedding-based retrieval, Flyder also runs a managed Vector DB. Upsert and delete documents directly through the API:

• GET /api/vectordb/collections — list your collections
• POST /api/vectordb/collections — create a collection, optionally uploading a file into it
• POST /api/vectordb/collections/{collection_id}/upload — upload a file into an existing collection
• DELETE /api/vectordb/collections/{collection_id} — delete a collection and all its data
• GET /api/vectordb/collections/{collection_id}/chunks — list the chunks in a collection, paginated
• POST /api/vectordb/upsert — chunk, embed, and upsert text into a collection
• DELETE /api/vectordb/delete — delete a document or a single chunk

Workflows & Chat Tasks

Beyond retrieval, the API runs the same automation building blocks available in the Flyder UI:

• GET /api/workflow/list / GET /api/workflow/details/{id} / POST /api/workflow/run/{id} — list, inspect, and trigger your saved workflows
• GET /api/tasks/chat / POST /api/tasks/chat/{task_id} — list and converse with chat-enabled AI tasks

---

Authentication

Every endpoint above accepts a Flyder API Key, generated from Settings > Flyder API Key once you're signed in. Send it as two headers:

Authorization: Bearer <your-api-key>
User-Email: [email protected]