Replay View
Watch an agent run live in the browser, or replay a finished one
praisonai replay view opens a local browser page that follows your agents as they run, or replays a run you recorded earlier.
from praisonaiagents import Agent
from praisonai.replay import record
with record(live=True) as rec:
Agent(name="writer", instructions="Draft a short blog post").start("AI agents")
print(f"praisonai replay view {rec.run_id}")Quick Start
- Install the UI
bash pip install "praisonai[ui]" - Run a YAML and watch it live
bash praisonai replay view agents.yamlThe YAML runs in the background, is recorded, and the browser follows it as it happens.
- Replay a finished run
bash praisonai replay list praisonai replay view run-1a2b3c4d5e6f
How It Works
| Target | What happens |
|---|---|
| (none) | Opens the list of recent runs |
run-1a2b3c4d5e6f | Replays that run |
./traces/run-1a2b3c4d5e6f.jsonl | Replays a trace file from any folder |
agents.yaml | Runs it, records it, and follows it live |
Options
| Option | Default | Description |
|---|---|---|
TARGET | recent runs | Run id, trace .jsonl path, or a YAML file to run and watch |
--live | false | Follow the run as it is written (implied for a YAML file) |
--port, -p | 8090 | Port for the replay UI |
--host | 127.0.0.1 | Host to bind to |
praisonai replay view run-1a2b3c4d5e6f --port 9000
praisonai replay view run-1a2b3c4d5e6f --liveThe viewer reads traces from the same place recording writes them — see Where Traces Live.
Best Practices
Record with live=True to follow from Python
record(live=True) writes every event immediately, so praisonai replay view RUN --live shows it as it happens.
Keep the default host
Stay on 127.0.0.1 and share a self-contained HTML export instead of exposing the server.