Explicit input modes
URL, feed item, raw text, Markdown, or pre-segmented blocks use distinct validated objects.
Summary API design
Feed Summary API Architecture
A stable API exposes what was received, what content was extracted, how the summary was structured, and which warnings or evidence apply.
REST APIOpenAPIasync jobsidempotencyevidence
API architecture
Contract-first summary resources
Separate source, extraction, summary, evidence, warnings, and processing state.
Structured output
Stable fields, null behavior, enums, length limits, evidence IDs, and coverage status.
Clear failures
Machine-readable extraction, validation, rate, policy, and processing error codes.
Job model
Synchronous when small, asynchronous when necessary
URL retrieval, PDFs, batches, and long content should not be forced through one short request.
202job accepted
IDEMPsafe client retries
STATEobservable processing
WEBHOOKoptional completion event
Safety + operations
Security begins at the URL boundary
A summarizer that fetches arbitrary URLs can become an internal-network proxy or uncontrolled crawler.
Fetcher controls
Restrict protocols, block private and link-local ranges, revalidate redirects, cap response sizes, enforce timeouts, and rate-limit by client and destination.
Process complex documents in constrained workers.
Output controls
Treat content as untrusted, validate model responses, escape rendered fields, avoid logging full content, and scope authentication by operation.
Document retention and deletion behavior.
Feed Signals
Detailed API design guide
Inputs, outputs, evidence, extraction states, async jobs, idempotency, errors, security, and tests.
Designing an Article Summarizer API: Inputs, Outputs, and Quality Controls
Design an article summarizer API with stable inputs, structured outputs, citations, length controls, and transparent failure modes.
Article SummarizationAPI DesignOpenAPI
Read guide