Conversation Workspace

Persist chats, tune personas, and monitor context usage.

No active context

Max messages

Max characters

Auto-compaction trims the oldest turns when limits are exceeded.

Persona preset

Select a preset to load curated instructions for the model.

Conversation History

Persistent context

Run your first prompt to see the conversation flow.

Model

Persona

Prompt

Include context

Streaming response

Context Insights

Auto-compaction removes the oldest turns when limits are exceeded.
Switch contexts to maintain separate conversations per user or topic.
Apply persona presets or save your own for quick prompt engineering.

Interaction Log

Interactions will appear here as you use the app.

Performance Metrics

Track latency and compare model behavior across your sessions.

Total requests

0

No conversations yet.

Average latency

0 ms

Latencies update after each response.

Fastest response

—

Awaiting data.

Slowest response

—

Awaiting data.

Model comparison

No model activity tracked yet.

Model	Requests	Avg latency (ms)	Fastest (ms)	Slowest (ms)	Median response size
No metrics yet. Send a prompt to populate this table.

Agent Configuration

Configure router/planner/executor models and loop behavior. Settings persist locally.

Router model

Router persona

Planner model

Planner persona

Executor model

Executor persona

Specialized model

Specialized persona

Max retries per loop

Expected finish JSON (minimal schema)

Not saved yet.

Agent user prompt

Run the agent to see router, planner, specialized and executor steps.

Agent v2 Workflow Builder

Drag nodes, connect stages, and export a JSON workflow.

Drag nodes here to start designing.

Agent v2 user prompt

Run the workflow to see node outputs and final answer.

Software Documentation

Project architecture, API endpoints, and usage guide.

Loading documentation…