What the Data for This Category Looks Like
The data sources for meeting minutes organization primarily consist of handwritten notes from in-person meetings, audio transcription files from online meetings, and agenda drafts submitted by attendees. Documents are generated either as a single file after each completed meeting, or as weekly batches of archived past meeting materials. A single document typically includes fields such as meeting topic, meeting time, attendee list, discussion topics, resolved matters, and follow-up items. Field units include integers (e.g., number of attendees), time units (e.g., minutes), and character units (e.g., length of agenda description). Document lengths vary widely, from hundreds to tens of thousands of characters. Some archived long documents contain complete records of multiple rounds of discussions.
What Constraints These Characteristics Impose on Source Citation and Traceability
The multi-source nature of meeting minutes requires the traceability function to associate original data in multiple formats, including handwritten notes, audio transcriptions, and agenda drafts, and accurately match source fragments corresponding to preset fields. The wide range of document lengths requires the traceability function to adapt to contexts of any length, preventing incomplete citations caused by long document truncation. The fixed field structure requires the traceability function to extract source content based on preset fields, and cannot flexibly adapt to free text without a fixed structure. The batch update rhythm requires the traceability function to efficiently traverse archived meeting documents, quickly locate target citation fragments, and maintain consistent traceability logic across all document batches.
Configuration Settings
Reference only; verify against the target FastGPT release and deployment documentation.
| Configuration Item | Recommended Setting | Rationale |
|---|---|---|
quoteMaxToken | 800–1200 token | The core cited content of a single meeting minute segment is typically 500–1000 tokens. This range covers complete citations for most core fragments. |
responseData | Includes sourceTitle, sourceText, sourceTimestamp | Matches metadata such as meeting minute document title, original text fragment, and timestamp to meet traceability requirements. |
detail | High | Meeting minutes require accurate association with original discussion details. High detail mode returns more granular original source fragments. |
| Single search limit | Top 6 entries | Covers the main discussion branches of a single meeting, avoiding excessive non-core redundant citations. |
| Chunk overlap length | 150 characters | Prevents cross-chunk discussion content in meeting minutes from being split, ensuring complete context for citation fragments. |
Three Common Configuration Errors
- The sourceText field appears empty in returned results. This occurs when the
responseDataparameter is not correctly configured, and the metadata item for returning original text fragments is not specified. - Citation fragments are excessively truncated. This occurs when the
quoteMaxTokenvalue is too small to cover the full length of the core citation segment of the meeting minute. - Association with specific timestamps of original meeting audio transcriptions fails. This occurs when the metadata field containing timestamps is not configured in
responseData, resulting in traceability only returning the document title without specific location information.
How to Verify Correct Configuration
- Use a standard meeting minute test file that includes multiple fields. Activate the citation traceability function, and check if returned results include original text fragments.
- Review returned metadata fields to confirm they include content specified by configuration items such as sourceTitle, sourceText, and sourceTimestamp.
- Adjust the
quoteMaxTokenparameter to different values, and observe whether the length of returned citation fragments changes accordingly. - Perform batch uploads of multiple archived meeting documents, and confirm the system returns associated source fragments according to configured rules.
The values provided here are common starting points, and should be evaluated against the reader’s own samples.
Question material comes from public community discussions. Configuration values are common starting points and should be measured against your own samples. Verified on 2026-09-28.