Commands
replay
Replay historical chat exports into ClawVault observations
The replay command ingests exported conversation history from supported platforms and processes it through the observation pipeline. This lets you backfill your vault with observations from sessions that occurred before ClawVault was installed.
Added in v2.2.0
Usage
clawvault replay --source <platform> --input <path> [options]Options
| Flag | Description | Default |
|---|---|---|
--source <platform> | Source platform (required) | — |
--input <path> | Export file or directory (required) | — |
--from <date> | Start date filter (YYYY-MM-DD) | — |
--to <date> | End date filter (YYYY-MM-DD) | — |
--dry-run | Preview replay without writing | false |
-v, --vault <path> | Vault path | Nearest vault |
Supported Sources
| Source | Expected Input |
|---|---|
chatgpt | Directory containing conversations.json |
claude | One or more .json export files |
opencode | .json or .jsonl session files |
openclaw | .jsonl session transcript files |
How It Works
- Normalize — Platform-specific parsers convert exports into a common message format
- Filter — Messages are filtered by the optional date range
- Group — Messages are bucketed by date
- Observe — Each day's messages are processed through the observation pipeline with a timestamp fixed to that date
- Reflect — A reflection pass runs across the full replay span
Replay requires an LLM API key configured in your environment for the observation compression step.
Examples
# Replay a ChatGPT export
clawvault replay --source chatgpt --input ~/chatgpt-export/
# Replay Claude conversations for January 2026
clawvault replay --source claude --input ~/claude-export/ --from 2026-01-01 --to 2026-01-31
# Dry run to see how many messages would be ingested
clawvault replay --source openclaw --input ~/.openclaw/sessions/ --dry-run
# Replay OpenCode sessions into a specific vault
clawvault replay --source opencode --input ~/opencode-logs/ -v ~/my-vault