snapshot-conversation
Save conversation progress to your Space with automatic AI enrichment and semantic indexing.
Overview
The snapshot-conversation tool is the primary way to save conversations and knowledge to your workspace. It automatically:
- •Generates vector embeddings for semantic search
- •Enriches content with AI-generated summaries
- •Extracts key points and concepts
- •Categorizes by topic (planning, implementation, research, etc.)
- •Tracks comprehensive metadata for analytics
When to Use
Call this tool:
- • Every 5-10 message exchanges during conversations
- • After making important decisions
- • When completing tasks or solving problems
- • At natural conversation milestones
- • Before context switching to preserve state
- • When you want to remember something for later
Think of this as "saving your progress" - use it liberally to build your knowledge base!
Parameters
contentREQUIREDType: string
The conversation content to snapshot. This can be a full conversation transcript, a specific discussion, meeting notes, or any text you want to save and search later.
workspaceIdOPTIONALType: string
The UUID of the workspace to save to. If omitted, the snapshot will be saved to your default workspace.
highlightsOPTIONALType: string[]
An array of user-specified highlights or important points. These help focus the AI enrichment process on what matters most to you.
["OAuth 2.0", "Supabase Auth", "GitHub provider"]privateOPTIONALType: boolean
Keep this snapshot private (only you can see). Default is false (shared with workspace team).
true- Private scope (only you)false- Workspace scope (team can see)
metadataOPTIONALType: object
Rich metadata object with 13 categories for comprehensive context capture. See the Metadata Specification for complete details.
Show metadata structure
{
client: {
name: "Claude Desktop",
version: "1.0.0",
platform: "darwin"
},
session: {
id: "session-123",
device: { type: "desktop", os: "macOS" }
},
conversation: {
threadId: "thread-456",
messageCount: 15
},
// ... 10 more categories
}Response
Returns a JSON object with the snapshot ID and status:
{
"success": true,
"snapshotId": "550e8400-e29b-41d4-a716-446655440000",
"message": "Snapshot saved to your Space. AI enrichment in progress."
}Examples
Basic Usage
snapshot-conversation({
content: "Discussed implementing OAuth 2.0 for user authentication.
Decided to use Supabase Auth for simplicity. Will set up GitHub
and Google providers next week."
})With Highlights
snapshot-conversation({
content: "Had a great brainstorming session about the new feature...",
highlights: [
"Real-time collaboration",
"WebSocket architecture",
"React concurrent features"
]
})Private Snapshot
snapshot-conversation({
content: "Personal notes: Need to refactor the auth module...",
private: true
})With Rich Metadata
snapshot-conversation({
content: "Implemented new React hooks for data fetching...",
highlights: ["useQuery hook", "caching strategy"],
metadata: {
client: {
name: "Claude Desktop",
version: "1.0.0",
platform: "darwin"
},
session: {
id: "session-abc123",
device: {
type: "desktop",
os: "macOS 14.0"
}
},
conversation: {
threadId: "thread-xyz789",
messageCount: 23,
durationMs: 1800000
},
workspace: {
path: "/Users/dev/my-project",
git: {
branch: "feature/data-fetching",
repo: "github.com/user/project"
}
},
content: {
codeLanguages: ["typescript", "tsx"],
topics: ["React", "data fetching", "hooks"]
}
}
})Best Practices
- ✓Snapshot regularly: Don't wait until the end - capture progress as you go
- ✓Use highlights: Help the AI focus on what's important to you
- ✓Include context: The more content, the better the AI enrichment
- ✓Send metadata: Rich metadata enables powerful analytics
- ✓Use private for personal notes: Keep sensitive information scoped appropriately