API
Run Kits and receive structured JSON
The Semarize API accepts conversation content plus a Kit code, runs the Kit, and returns typed Brick results with reasons, confidence, evidence, and version metadata.
Request
{
"kit_code": "6AD3F0B6",
"input": { transcript }
}
Response
{
"run_id": "run_01J...",
"kit_version_id": "kv_2.3",
"output": { bricks }
}
Definition
An API for conversation signals
This page explains the product-level API response contract: what you send, what comes back, and how the response fits into systems. For endpoint details, use the developer docs.
Conversation in
Send transcripts, emails, chats, or notes as JSON input. Speaker attribution and metadata can be included when available.
Kit selected
Choose the evaluation framework by passing a Kit code. The Kit decides which Bricks run against the conversation.
Structured output
Receive output.bricks with typed values, reasons, confidence, evidence, and version identifiers.
Works with webhooks
Trigger a run when a transcript is ready, poll on a schedule, or push conversation content directly from your stack.
Lands in systems
Use the output in CRM enrichment, BI tables, QA queues, automation tools, and internal applications.
Production controls
Authenticate with API keys and keep downstream integrations pinned to stable Kit and Brick versions.
Request and response
Content in. Signals out.
A run is the execution of a Kit against one conversation input. The same Kit version returns the same response shape on every run.
Send input
Pass the conversation content as JSON. It can come from a meeting recorder, note-taker, dialler, chat system, email thread, or your own app.
Pass Kit code
Use the Kit code to choose the framework. Optional version pinning can keep integrations stable during framework changes.
Run Bricks
Semarize evaluates every Brick in the Kit, using attached knowledge bases when the Kit is grounded.
Return JSON
The response includes run metadata and structured Brick results that systems can parse without reading a summary.
Response contract
The fields systems care about
The API response is shaped for integration work: stable identifiers, typed values, evidence, and enough metadata to audit what happened.
Run metadata
Identifiers and state for the run, used for storage, debugging, retries, and audit trails.
Kit metadata
The exact framework version that produced the result, so historical output stays explainable.
Brick values
The structured outputs your CRM, warehouse, workflow tool, or internal app can consume.
Evidence spans
The source text behind the value, useful for QA, review queues, manager coaching, and audit questions.
Sync or async patterns
Use synchronous runs for interactive flows and asynchronous runs when transcript processing is part of a larger pipeline.
Downstream mapping
Map each Brick result once, then reuse the same output shape in Salesforce, HubSpot, Snowflake, BigQuery, Zapier, Make, or n8n.
Positioning
API output not another dashboard
Semarize is not trying to become the place your team logs into to read every call. It is the layer that returns reliable conversation data to the systems you already use.
Dashboard-first platform
Semarize API
Reliability
Designed for production workflows
A conversation intelligence API has to behave like infrastructure. That means predictable schemas, auth, versioning, and audit trails.
Versioned evaluation
Kit and Brick version identifiers let teams improve evaluation logic without silently changing the meaning of old results.
Authenticated access
API usage is authenticated with keys so systems can process conversations without user dashboard sessions.
Evidence-backed output
Values can include the source evidence and reasoning needed for review, audit, or human escalation.
Ready to build
Send one transcript. Get structured fields back.
Use the API when conversation data needs to land in your own product, CRM, warehouse, or automation stack.
No card required. Start testing in minutes.
Questions
Frequently asked questions
Short answers about API inputs, outputs, and how the Semarize run contract relates to Bricks and Kits.
From the blog
Further reading
Practical pieces for teams evaluating API-first conversation intelligence, structured JSON, and API integration patterns.