Maintaining a personal knowledge base is useful for developers and researchers alike. Over time, interactions with AI chat platforms and agent sessions accumulate, often residing in dormant JSON files. The new conversation-archive AI agent skill addresses this directly by converting these scattered histories into a structured, searchable format. It's designed for anyone who regularly uses AI chats and wants to integrate that historical data into their active knowledge base, moving beyond simple storage to actual utility.
This tool is specifically built to import your AI chat exports from platforms like ChatGPT, Claude, and Perplexity, as well as your agent session logs. Instead of keeping this information as raw data, conversation-archive transforms it into dated markdown pages stored under a conversations/ directory. This process turns years of past interactions into content that is not only searchable but also includes fact-extracted details. The core idea is to make your personal AI interaction history an accessible resource.
From Export to Knowledge Base
The import process within conversation-archive is thorough. It handles the conversion of your chat exports and agent session logs into individually dated markdown pages. Before writing anything to your knowledge base, the tool validates the format of the incoming data, ensuring consistency and integrity. A critical step involves redacting secrets and personally identifiable information (PII) to maintain your privacy and security. The skill also features mechanisms for backfilling any detected gaps in your historical data, aiming for a complete record. This two-part approach ensures that your raw chat history becomes structured, clean, and secure content within your personal knowledge base.
Core Features for Data Integrity
conversation-archive provides several focused features to manage your data effectively. It supports native ingestion from six different export formats, offering flexibility regardless of your primary AI chat platform. A mandatory secret scanning process runs before any data is written, acting as a safeguard for sensitive information. To prevent conflicts, especially when multiple untitled threads occur on the same day, the skill uses thread hashes for collision prevention. Furthermore, automatic gap detection and healing ensure the continuity and completeness of your archive. During the import, it also performs fact extraction, turning conversation details into actionable data points rather than just raw text.
Practical Retrieval and Search
Once your data is archived, the retrieval capabilities become useful. The skill enables specific timeline queries, allowing you to trace discussions over time. You can perform thread tracing to follow the evolution of a particular conversation. For example, if you ask 'when did I first discuss large language models', the tool searches chronologically through your archived conversations to provide a precise answer. This functionality transforms a collection of past interactions into a dynamic resource for recall and historical context, supporting focused research or development tasks by surfacing relevant information from your personal history.
How to Engage the Skill
You can activate conversation-archive using several direct triggers. Common commands include 'ChatGPT export' or 'Claude export' when you have new data to import. For agent session logs, you can use 'archive my session transcripts'. If you need to find specific historical information, a trigger like 'when did I first discuss [topic]' initiates a chronological search through your archives. It's important to note what this skill is not designed for: it does not handle audio processing (that's for voice-note-ingest), human meeting transcripts (use meeting-ingestion), capturing single items (use capture), or broad idea tracing across your entire knowledge graph (that's concept-synthesis). conversation-archive focuses specifically on organizing and making accessible your AI conversation history.
FAQ
Q: What chat platforms does conversation-archive support? A: It supports exports from platforms like ChatGPT, Claude, and Perplexity, in addition to agent session logs, with native ingestion for six different export formats.
Q: How does the skill handle sensitive information? A: It includes mandatory secret scanning and PII redaction as part of its import process, ensuring sensitive data is protected before being written to your knowledge base.
Q: What if I have many untitled chats from the same day? A: The skill employs collision prevention logic using thread hashes to uniquely identify and organize same-day untitled threads.
This skill makes your past AI interactions a dynamic part of your active knowledge. It enables quick retrieval of facts and context from your personal history.
