Skip to content

pChronicle

pChronicle logo

Chronicled Experience for the Agent Era

makes every agent run easier to understand and improve

Agent experience is the sum of everything an agent did. pChronicle is an Agent trajectory storage engine: it records that experience at the unit that matters — the Run — and makes every Run easier to understand and improve. Use it to browse, query, exchange, and serve run Datasets produced by Persisting or by supported external formats.

In Persisting, pChronicle stores and queries trajectory history. It can run as a local tool or be deployed as a service in front of many paths.

What you will complete

The first walkthrough creates temporary data, opens it, runs a read-only summary, and answers one SQL question. You do not need a production store first.

The one object you work with

A Dataset is a path: a local directory or file, or an object-store URI prefix. pChronicle discovers and normalizes the supported data inside that path. Dataset pins (@name) are locators; after resolution the engine only sees the path.

A Dataset can be written as:

  • a local directory or file (./local/path);
  • an object-store URI prefix (s3://bucket/prefix);
  • a dataset pin that resolves to either location (@name).

pChronicle discovers and normalizes the supported data inside that path. You do not need to understand its internal files or storage layout before using the CLI.

Start here

Try the built-in walkthrough without preparing any data:

pchronicle onboard

Or inspect and query an existing Dataset:

pchronicle list ./trajectory-data
pchronicle stats overview ./trajectory-data
pchronicle query ./trajectory-data \
  --sql 'SELECT COUNT(*) AS runs FROM dataset.runs'

Choose your next step

Start with Explore your first Dataset if you are learning the command line. It creates temporary data and lets you complete a read-only query before you connect a real source.

When you already have a question, follow the matching path:

  • Inspect a Dataset: pchronicle list DATASET or pchronicle stats DATASET
  • Run a common report: pchronicle stats overview DATASET
  • Ask a custom SQL question: pchronicle query DATASET --sql SQL
  • Name a Dataset: pchronicle dataset pin NAME DATASET
  • Import or export runs: Exchange data
  • Analyze with an Agent: pchronicle agent codex DATASET
  • Open the local UI and API: Serve a Dataset

A useful reading order

  1. Explore your first Dataset to complete a read-only query.
  2. Discover and query when you are ready to inspect real data.
  3. Exchange data when a supported external format is involved.
  4. Serve a Dataset when another tool needs local access.
  5. CLI reference when you need exact flags, budgets, or output formats.

Keep reading