url-ingest Backends
Processing and storage backends are pluggable HTTP services. Use the platform defaults, or deploy your own and pay only the minimum circulation fee.
Backend types
- Processing — given a URL, produces structured markdown + images + raw credits. Examples: glm-ocr (PDFs), youtube-transcript (YouTube), MinerU-OCR.
- Storage — persists the processed content, serves progressive reads + semantic search. Currently one platform implementation: ctx-storage.
Ownership
- Platform — registered by an admin, visible to every project as a default option.
- Project — registered by a project owner, visible only to that project. Pays min_circulation_fee per billing event regardless of usage.
Namespace slugs
Each backend gets a stable namespace_id used as the cost-cache key. You pick a slug; the platform composes the full id:
- Platform:
ctxkit.<backend_type>.<slug> - Project:
proj.<project_id>.<backend_type>.<slug>
Pick a slug that survives redeployments (include a version suffix like -v1). Re-registering with the same slug preserves billing continuity; changing the slug resets the cost cache.
Implementing your own
Your backend speaks the same HTTP + HMAC contract the platform backends use. The url-ingest project config page has a detailed Implement tab with the full protocol, signing helpers, and a FastAPI scaffold.