llm-browser-multicast-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| COUNCIL_CONFIG | No | Path to the council configuration JSON file. If not set, searches for council.config.json in current directory or ~/.config/llm-council/config.json |
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 | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| council_discussA | Run a council discussion: all models answer in parallel, peer-review each other anonymously, then a chairman synthesizes the best answer. |
| roundtable_discussA | Run a roundtable discussion: models take turns responding, building on previous responses. More natural, conversational debate style. |
| debate_discussA | Run a structured debate: models state positions, discuss in rounds, restate final positions, then a chairman synthesizes points of agreement and disagreement. |
| list_modelsA | List all available models from configured providers. |
| ask_modelA | Ask a single model a question. Use this for interactive, facilitator-style discussions where Claude Code orchestrates the conversation and the user can intervene between rounds. |
| ask_modelsA | Ask multiple models the same question in parallel. Returns all responses at once. Use for gathering diverse opinions that Claude Code can then synthesize with the user. |
| deep_researchA | Run ChatGPT Deep Research on a query. Performs in-depth web research using ChatGPT's Deep Research mode (5-15 min). Returns comprehensive results with sources. Note: blocks other ChatGPT operations during research. |
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 7 tools
Each tool serves a clearly distinct purpose: three different discussion formats (council, roundtable, debate) are explicitly differentiated by their mechanics, ask_model vs ask_models distinguishes single vs. parallel queries, and list_models and deep_research are unique. There is no meaningful overlap that would cause an agent to misselect.
Most tools follow a verb_noun pattern (list_models, ask_models, ask_model) but the three discussion tools use a noun_verb pattern (council_discuss, roundtable_discuss, debate_discuss) and deep_research breaks the convention entirely. Despite the mixing, the names are readable and understandable.
With 7 tools, the set is well-scoped for a multi-LLM discussion and query server. Each tool covers a distinct mode of interaction, and there is no bloat or trivial filler.
The tool surface completely covers the domain of orchestrating multi-LLM discussions: listing available models, querying one or many models, running three distinct discussion formats, and performing deep research. There are no obvious missing operations that would leave agents stuck.