Product replay

See a source become a briefing.

This page replays a bounded excerpt from a sample generated by the real Perseus Context Engine 1.0.27 source candidate. 1.0.27 is not the published package release. The compatibility memory pointer is excluded. It does not run a model, call an API, measure your tokens, or create a production receipt.

1. Source2. Resolve3. Review

Inputdemo/sample-context.md

@perseus v1.0.27

# Sample workspace briefing

Rendered: @date format="YYYY-MM-DD HH:mm z"

## Repository

- Name: sample-api
- Branch: main
- API port: 3001
- Test status: fixture value only
- Last decision: review the webhook integration before release

## Boundary

This sample contains no secrets, customer data, network calls, or shell directives.

ArtifactAGENTS.md

No artifact loaded.

Activate “Replay committed render” to load the generated output and its SHA-256 metadata.

Idle. No network request has run.

What this replay does and does not show

The interaction loads the committed bounded excerpt and its metadata. The excerpt comes from a local Perseus Context Engine render and excludes the compatibility memory pointer. It does not prove model behavior, memory retrieval, token savings, customer performance, or a live service.

Run it locally

Build a complete briefing in your own workspace.

Author sources, render the briefing, inspect the output, then give the file to the assistant you already use.

pip install perseus-ctx==1.0.26
perseus quickstart
perseus render .perseus/context.md -o AGENTS.md