prodius-mcp
OfficialServer Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| PRODIUS_BASE_URL | No | Prodius local server URL. Default: discovered from settings (http://127.0.0.1:5555). | |
| PRODIUS_MCP_DEBUG | No | Set to 1 to log the text of voice messages to stderr (otherwise only their length). Default: off. | |
| PRODIUS_MCP_TOKEN | No | Auth token for the local Prodius server. Default: read from the Prodius app's settings.json. | |
| PRODIUS_AGENT_NAME | No | Label shown so you know which agent is asking. Default: Unknown Agent. | |
| PRODIUS_HOOK_DEBUG | No | Set to 1 to let prodius-permission-hook log commands and answers, and write a log file. Default: off. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {} |
| experimental | {
"claude/channel": {}
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| ask_userA | Ask the user a question via voice overlay. Sub-agents must NOT call this — return results to your parent agent. Plain text, one sentence, ~80 chars (100 max). |
| notifyA | Show the user a short message in a desktop overlay, without waiting for a reply. Use it as often as the user's own instructions say; with no such instructions, only for results they need to see while away from the terminal. Sub-agents must NOT call this. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 2 tools
The two tools have clearly distinct semantics: ask_user blocks for a reply, while notify sends a non-blocking message. An agent can easily choose between them based on whether user input is required.
Both names are lowercase and readable, but ask_user follows a verb_noun pattern while notify is a bare verb. The minor inconsistency does not create confusion.
Two tools is thin, but the server's apparent scope is limited to user overlay communication. Each tool earns its place, though the surface feels minimal.
The core interaction modes—asking for input and sending notifications—are covered. Minor gaps might include confirmation-specific helpers, but agents can work around them.