roku-debug-mcp
Provides AI agents with Roku BrightScript debugging tools, including log reading, scene graph inspection, breakpoint management, variable inspection, and execution control.
README
roku-debug-mcp
MCP server giving AI agents the full VS Code Roku debug experience.
Exposes Roku's BrightScript debugging capabilities as MCP tools so AI agents can read logs, inspect the scene graph, step through code, read variables, and set breakpoints — the same info a developer sees in the VS Code Roku extension.
Architecture
graph TB
subgraph "AI Agent (Hermes, VS Code, etc.)"
MCP[<b>MCP Client</b><br/>stdio JSON-RPC]
end
subgraph "roku-debug-mcp (MCP Server)"
Server[<b>MCP Server</b><br/>21 tools]
Config[<b>Config</b><br/>ROKU_* env vars]
Server --> Config
end
subgraph "Roku Device"
direction LR
subgraph "Port 80 — HTTP"
Installer[<b>Sideloader</b><br/>Digest auth<br/>Expect: 100-continue]
end
subgraph "Port 8060 — ECP"
ECP[<b>ECP Client</b><br/>Device info<br/>Scene graph<br/>Postback/keys]
end
subgraph "Port 8081 — Binary Debug"
Debug[<b>Debug Client</b><br/>Binary protocol<br/>BSDBG magic]
end
subgraph "Port 8085 — Telnet"
Console[<b>Text Console</b><br/>Fallback logs]
end
end
MCP --> Server
Server --> Installer
Server --> ECP
Server --> Debug
Server --> Console
Protocol Layers
| Port | Protocol | Auth | Purpose |
|---|---|---|---|
| 80 | HTTP | Digest + Expect: 100-continue | Channel sideloading |
| 8060 | ECP HTTP | None | Device info, scene graph, screenshots |
| 8081 | Binary | None | Primary debug protocol (VS Code uses this) |
| 8085 | Telnet | None | Text console (fallback) |
Binary Debug Protocol (Port 8081)
sequenceDiagram
participant C as Client (roku-debug-mcp)
participant R as Roku Device (port 8081)
C->>R: Handshake<br/>[magic(8)][protocol_version(4)]
R-->>C: [magic(8)][protocol_version(4)][packet_len(4)][revision]
Note over C,R: Request/Response Format:<br/>[packet_length(4)][request_id(4)][cmd_code(4)][payload]
C->>R: GET_THREADS (cmd=3)
R-->>C: THREADS response
C->>R: STACKTRACE (cmd=4, thread_index)
R-->>C: Stack frames
C->>R: ADD_BREAKPOINTS (cmd=7)
R-->>C: Confirmation
Note over C,R: Update notifications (request_id=0):<br/>CONNECT_IO_PORT, ALL_THREADS_STOPPED, etc.
Handshake Magic: 0x0067756265647362 (b"bsdebug\0" little-endian)
Sideloading Flow (Port 80)
sequenceDiagram
participant C as Client
participant R as Roku (port 80)
C->>R: POST /plugin_package (Expect: 100-continue)
R-->>C: 401 Unauthorized (WWW-Authenticate: Digest)
C->>C: Compute digest hash
C->>R: POST /plugin_package (Authorization: Digest)
R-->>C: 100 Continue
C->>R: [ZIP payload]
R-->>C: 200 OK [chunked response with Dev Kit HTML]
What AI Can Do With This
- Read device info — model, version, app currently running
- Inspect scene graph — the full node hierarchy of the running app
- Read console logs — stdout from the running BrightScript channel
- List threads — see all execution threads and their stop states
- Read stack traces — frame-by-frame call stack for any stopped thread
- Inspect variables — locals, globals, and scene graph component state
- Execute code — run arbitrary BrightScript in a stopped frame
- Manage breakpoints — add, list, remove breakpoints by file/line
- Step execution — step over, step into, step out, or continue
- Sideload channels — upload and install test channels with remote debug
Quick Start
1. Install
cd /home/dom/src/roku-debug-mcp
pip install -e .
2. Configure Environment
export ROKU_DEVICE_IP=192.168.1.10 # Roku device IP
export ROKU_DEV_USER=rokudev # Dev channel username
export ROKU_DEV_PASSWORD=your-password # Dev channel password
3. Register in Hermes
Add to ~/.hermes/mcp-servers.json:
{
"roku-debug-mcp": {
"command": "roku-debug-mcp",
"args": []
}
}
4. Use in a Hermes Session
The AI agent will now have access to 21 new tools:
roku_device_info()
roku_scene_graph()
roku_debug_threads()
roku_debug_stacktrace(thread_index=0)
roku_debug_variables(thread_index=0, frame_index=0)
roku_debug_execute(thread_index=0, frame_index=0, code="x = 42")
roku_debug_breakpoints_add(breakpoints=[{...}])
roku_debug_console_output()
Available Tools
Device / UI Tools (ECP — port 8060)
| Tool | Description |
|---|---|
roku_device_info |
Device model, version, etc. |
roku_current_app |
Currently running app |
roku_scene_graph |
Full scene graph node hierarchy |
roku_postback |
Send postback to channel |
roku_launch_uri |
Launch a URI |
roku_key |
Send remote control key |
roku_screenshot |
Capture screen image |
Debug Tools (Binary Protocol — port 8081)
| Tool | Description |
|---|---|
roku_debug_threads |
List all threads |
roku_debug_stacktrace |
Get stack frames |
roku_debug_variables |
Read variables in a frame |
roku_debug_execute |
Run BrightScript code |
roku_debug_breakpoints_add |
Add breakpoints |
roku_debug_breakpoints_list |
List active breakpoints |
roku_debug_breakpoints_remove |
Remove specific breakpoints |
roku_debug_breakpoints_remove_all |
Clear all breakpoints |
roku_debug_continue |
Resume execution |
roku_debug_step |
Step execution |
roku_debug_stop |
Pause execution |
roku_debug_console_output |
Get stdout lines |
roku_debug_protocol_info |
Debug protocol version |
Installer Tools (HTTP — port 80)
| Tool | Description |
|---|---|
roku_install |
Sideload a channel ZIP |
roku_launch_remote_debug |
Launch with remote debugging enabled |
Testing
Unit Tests (Mock Roku Server)
# Run all tests (uses mock server on ephemeral ports)
pytest tests/ -v
# Mock server runs automatically via conftest fixtures
# No manual setup required
Integration Tests (Real Roku Device)
# Requires env vars set
ROKU_DEV_IP=10.71.71.151 \
ROKU_DEV_PASSWORD=your-password \
pytest tests/test_integration_real_device.py -v
CI/CD
- Unit tests run on GitHub Actions Ubuntu runners
- Integration tests run on self-hosted runner (10.71.71.90) with LAN access to real Roku
Project Structure
src/rokumcp/
config.py # Environment-based configuration
protocol.py # Binary protocol constants and Stream I/O
debug_client.py # Synchronous binary debug client (port 8081)
text_console.py # Telnet text console client (port 8085)
ecp.py # ECP HTTP client (port 8060)
installer.py # HTTP Digest-auth sideloader (port 80)
server.py # MCP server entrypoint — 21 tools
tests/
conftest.py # Pytest fixtures (mock server setup)
mock_roku_server.py # Mock Roku device simulator
test_protocol.py # Stream round-trips, ProtocolVersion
test_config.py # Config defaults, from_env
test_ecp.py # ECP HTTP client
test_text_console.py # Telnet console client
test_installer.py # Digest auth + multipart
test_debug_client.py # Full E2E vs mock binary server
test_integration_real_device.py # Real device (gated on env vars)
fixtures/ # Test channel ZIP fixtures
Protocol Reference
Implementation derived from Roku's official reference:
See AGENTS.md for the full protocol specification and wire formats.
Development
Debugging the Protocol
# Run mock server manually
python tests/mock_roku_server.py
# Test specific protocol interaction
ROKU_DEVICE_IP=127.0.0.1 ROKU_DEBUG_PORT=8081 python -m rokumcp.server
Building
pip install -e .
roku-debug-mcp # runs MCP server over stdio
License
Apache-2.0
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.