# Fast mode Fast mode supports synchronous summarization for inline conversation text with immediate results after processing completes. ## Capabilities - Single-call transcript summarization - Immediate summarization results - Optimized for bounded conversations and meeting content ## Input The API accepts the following input: - Supports conversation or transcript text from any system - Limits input size to 96 KB ## Output The API returns the summarization result as human-readable rendered text in the `result.text` field. The content of `result.text` depends on the `task` you requested: | Task | Content of `result.text` | | -------------- | ------------------------------------------------------------------------------------------------------------------------------------------------ | | `recap` | A `Recap:` section with a single concise sentence or short paragraph. | | `summary` | A `Summary:` section with a detailed paragraph or multi-paragraph summary. | | `action_items` | An `Action Items:` section grouped by owner. Each owner appears as a `**Name**` markdown header, followed by their items as `- ` bulleted lines. | | `full_summary` | All three sections — `Recap:`, `Summary:`, and `Action Items:` — concatenated with blank lines between them. | For example, a `full_summary` response returns `result.text` similar to: ```text Recap: The team agreed to delay the launch by one week. Summary: The discussion focused on finalizing the project timeline. Engineering requested additional time for validation, and the team aligned on shifting the launch schedule accordingly. Action Items: Engineering - Complete validation before the new launch date. Marketing - Adjust the campaign timeline to match the revised launch. ``` ## Summarization options These options control the summarization task and output language. | Parameter | Type | Default | Description | | -------------- | ------ | ---------------- | ----------------------------------------------------------------------------------------------------------------------------------- | | `summary_type` | string | `"conversation"` | The type of content being summarized. Currently only `"conversation"` is supported. | | `task` | string | `"full_summary"` | The summarization task. Accepted values are `"recap"`, `"action_items"`, `"summary"`, and `"full_summary"`. | | `language` | string | `"en-US"` | Output language as a BCP-47 locale code. Must be one of the [supported locales](/docs/ai-services/summarizer/#supported-languages). | You can optionally include a `reference_id` to track the request in your own systems. ## Summarize text Use this endpoint to send inline transcript text for synchronous summarization: **`POST /aiservices/summarizer/summarize`** ### Example request (cURL) This cURL example sends conversation text to the Summarizer API and returns a rendered summary: ```shell curl -X POST https://api.zoom.us/v2/aiservices/summarizer/summarize \ -H "Authorization: Bearer $TOKEN" \ -H "Content-Type: application/json" \ -d '{ "input": { "text": "Speaker A: Let'\''s finalize the timeline. Speaker B: We need one more week." }, "config": { "summary_type": "conversation", "task": "summary", "language": "en-US" }, "reference_id": "meeting-789" }' ``` #### Response (200) ```json { "request_id": "req_123", "task": "summary", "result": { "text": "Summary:\nThe discussion focused on finalizing the project timeline, with agreement to extend it by one week to allow additional validation." }, "model": "zoom-summarizer-v1" } ```