For AI agents: a documentation index is available at the root level at /llms.txt and /llms-full.txt. Append /llms.txt to any URL for a page-level index, or .md for the markdown version of any page.
  • Workflows
    • Overview
  • Get Started
    • MCP Quick Start
    • API Quick Start
  • MCP Server
    • Overview
    • Installation
      • Overview
      • get_all_minds
      • search_minds
      • create_mind
      • get_signed_upload_url
      • upload_file_to_signed_url
      • create_snapshot
      • get_snapshot_status
      • simulate
      • gong_list_calls
      • gong_get_transcript
      • gong_get_call_details
      • fathom_list_meetings
      • fathom_get_transcript
      • fathom_get_meeting_details
      • fireflies_list_transcripts
      • fireflies_get_transcript
      • fireflies_search_transcripts
  • Direct API
    • Overview
    • Authentication
    • Error Handling
  • Core Concepts
    • Minds, Snapshots, & Simulations
    • SOC 2, HIPAA, & GDPR
LogoLogo
LogoLogo
On this page
  • Parameters
  • mindId
  • digitalTwinId
  • artifactId
  • Request
  • Response
  • Response Properties
  • Processing Time Estimates
  • What the AI Learns
  • Error Responses
  • Next Steps
  • See Also
MCP ServerTools Reference

create_snapshot

Train an AI model from conversation data
Was this page helpful?
Previous

get_snapshot_status

Check AI training progress
Next
Built with
CREATE

create_snapshot

Starts AI training to analyze transcript data and learn communication patterns, decision-making style, tone, vocabulary, and personality traits. This is an asynchronous operation that typically takes 5-15 minutes.

The create_snapshot tool processes uploaded transcript data into a trained AI model. The result is a snapshot—a point-in-time representation of how a person thinks and communicates.

Parameters

mindId

Typestring
RequiredYes
DescriptionThe mind ID from the create_mind response

digitalTwinId

Typestring
RequiredYes
DescriptionThe digital twin ID from the create_mind response

artifactId

Typestring
RequiredYes
DescriptionThe artifact ID from the get_signed_upload_url response

Request

1{
2 "mindId": "abc-123-def-456",
3 "digitalTwinId": "xyz-789-uvw-012",
4 "artifactId": "artifact-456-def-789"
5}

Response

1{
2 "message": "Memory bank snapshot processing initiated",
3 "mindAssessmentId": "123e4567-e89b-12d3-a456-426614174003",
4 "artifactId": "123e4567-e89b-12d3-a456-426614174004"
5}

Response Properties

PropertyTypeDescription
messagestringStatus message confirming processing has started
mindAssessmentIdstringRequired for status checks - Snapshot identifier (also called snapshotId)
artifactIdstringAssociated artifact identifier (matches input parameter)

Your AI model is now training in the background. Use get_snapshot_status to check when it’s ready for simulations.

Processing Time Estimates

File SizeConversationsTypical Time
< 1 MB10-203-5 minutes
1-10 MB30-1005-15 minutes
10-50 MB100-50015-30 minutes
50-100 MB500+30-45 minutes

Factors affecting processing time include file size, number of unique speakers, conversation complexity, and current system load.

What the AI Learns

During processing, Mind Reasoner’s AI analyzes your transcript to extract:

Communication Style
  • Vocabulary and phrasing
  • Sentence structure
  • Formality level
  • Speech patterns
Emotional Patterns
  • Tone and mood
  • Empathy expression
  • Confidence levels
  • Reaction patterns
Decision-Making
  • Problem-solving approach
  • Question handling
  • Prioritization style
  • Reasoning patterns

Error Responses

404 Not Found
1{
2 "error": "Not Found",
3 "message": "Artifact with ID 'artifact-456' does not exist"
4}

Solution: Verify the artifactId is correct and ensure the file upload completed successfully.

400 Bad Request
1{
2 "error": "Bad Request",
3 "message": "Invalid file format or corrupted data"
4}

Solution: Verify the file is a valid .vtt, .pdf, or .docx format and ensure the file contains actual conversation data. Try re-uploading if needed.

409 Conflict
1{
2 "error": "Conflict",
3 "message": "A snapshot is already processing for this mind"
4}

Solution: Wait for the current snapshot to complete. Check status with get_snapshot_status. You can only process one snapshot at a time per mind.

Next Steps

1

Check Status

get_snapshot_status - Poll to check when training is complete

Your AI assistant can continue other work while the snapshot processes. It will automatically check the status and notify you when training is complete.

See Also

upload_file_to_signed_url

Previous step: Upload your file

get_snapshot_status

Next step: Check training status