Parse, split, and organize exported AI conversations into your knowledge workspace.
Ingestion surface for exported AI chats and related conversational records. Turn months of AI conversations into structured, searchable, graph-linked knowledge assets.
| Ingest | Split | Reconstruct |
|---|---|---|
| Import exported AI conversations from ChatGPT, Claude, and other platforms. | Automatically separate messages by speaker, topic, and conversation thread. | Rebuild threaded conversations with full context and speaker attribution. |
Batch import of conversation exports in JSON, Markdown, and HTML formats. The queue validates file formats, extracts metadata, and prepares conversations for parsing — no manual cleanup needed.
Learn more about Import Queue →
Automatic separation of messages by speaker role, topic shift, and thread boundary. The splitter identifies where conversations change direction and creates clean, navigable segments.
Learn more about Message Splitter →
Identify and label participants across conversations — human users, AI assistants, system messages. Role mapping enables filtered views ("show me only the AI's code suggestions") and attribution analytics.
Browse parsed conversations as reconstructed threads with full context — timestamps, speaker attribution, topic labels, and linked entity references displayed inline.
AI platformexport data as flat JSON or Markdown files. Without parsing, conversations are unsearchable walls of text. The Conversation Parser transforms exports into indexed, navigable records.
A 200-message AI conversation might contain 10 genuinely valuable insights surrounded by iteration, corrections, and small talk. Message splitting and topic segmentation surface the signal.
An insight from a Claude conversation three months ago connects to a ChatGPT thread from last week — but you'd never know it. Parsed conversations feed into the Knowledge Graph, linking concepts across sources.