pyagent-studio 0.1.0

Terminal-based interactive workbench for designing, simulating, and debugging agent systems

pyagent-studio

Terminal-based interactive workbench for designing, simulating, debugging, and governing multi-agent LLM systems. Built with Textual for a rich TUI experience.

Install

pip install pyagent-studio[tui]    # Full TUI (requires textual)
pip install pyagent-studio          # Services only (no TUI)

Depends on: pyagent-blueprint, pyagent-trace, click.

Quick Start

# Launch with a blueprint
pyagent-studio blueprint.yaml

# Or launch and browse discovered blueprints
pyagent-studio

Screens

Dashboard

List discovered YAML blueprints, view validation status, agent/workflow counts.

┌─────────────────────────────────────────────────────────────┐
│  PyAgent Studio — Dashboard                                 │
├─────────────────────────────────────────────────────────────┤
│ File                  │ Name             │ Agents │ Status   │
│ customer_support.yaml │ customer-support │ 3      │ ✓ Valid  │
│ research_agent.yaml   │ research-agent   │ 2      │ ✓ Valid  │
└─────────────────────────────────────────────────────────────┘

Editor

YAML editing with live validation panel showing Pydantic errors as you type.

Graph

ASCII-rendered workflow DAG showing agents as boxes and edges as pattern connections.

Simulation

Run blueprints with MockLLM (no API keys needed), stream results with timing and token counts.

Traces

Browse recorded trace spans from pyagent-trace Recorder JSONL files.

Cost

Cost breakdown tables by pattern, agent, and model with bar chart visualization.

Governance

Validation issues, compliance scoring (% of checks passing), and semantic diff between blueprint versions.

Key Bindings

Key	Action
Tab / Shift+Tab	Switch screens
q	Quit
Ctrl+S	Save blueprint
Ctrl+V	Validate
Ctrl+R	Render graph
Ctrl+T	Run simulation
Ctrl+D	Show diff

Services (Headless API)

Use the services layer without the TUI for scripting and CI:

from pyagent_studio import BlueprintService, SimulationService, GovernanceService

# Load and validate
svc = BlueprintService()
spec = svc.load("blueprint.yaml")
issues = svc.validate()
print(svc.summary())

# Run simulation
import asyncio
sim = SimulationService()
result = asyncio.run(sim.run(spec, "support", "I can't see my invoice"))
print(result.output)

# Governance
gov = GovernanceService()
report = gov.check_compliance(spec)
print(gov.format_report(report))

TraceService

from pyagent_studio import TraceService

traces = TraceService()
spans = traces.load("traces.jsonl")
llm_calls = traces.query(event_type="llm_call")
print(traces.summary())

Full Documentation

See pyagent.dev for full API reference and integration guides.