tg-recall-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| TG_RECALL_HOME | No | Override the home directory for configuration and data. Precedence: --home, TG_RECALL_HOME, then system defaults. | |
| TG_RECALL_PROFILE | No | Override the active profile. Precedence: --profile, TG_RECALL_PROFILE, configured active profile, then default. |
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 | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| list_allowed_chatsA | List cached chats allowed by AI access policy. |
| list_scopesA | List saved sync scopes. |
| search_messagesC | Search indexed messages and transcripts with citations. |
| get_message_contextC | Return nearby cached messages around a cited message. |
| ask_archiveC | Answer using bounded local archive retrieval. |
| retrieve_evidenceC | Return bounded cited local evidence with optional genuine local-vector retrieval. Read-only; never builds or changes an index. |
| query_knowledge_catalogB | Read compact cited catalog metadata in one exact saved scope; no raw source body or writes. |
| inspect_research_sessionC | Read compact checkpoint/session metadata for an explicitly allowed session; no mutation. |
| expand_cited_sourcesC | Expand only explicit cited Telegram sources under current scope and bounded payload/work budgets. |
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 9 tools
Most tools are distinct in purpose: search/research tools (search_messages, retrieve_evidence, ask_archive, query_knowledge_catalog, inspect_research_session) vs. retrieval/expansion (get_message_context, expand_cited_sources) vs. admin listing (list_allowed_chats, list_scopes). However, there is notable overlap between search_messages, retrieve_evidence, ask_archive, and query_knowledge_catalog — all involve retrieving or searching cited/local content, making it ambiguous which to select for a research query.
Names follow a loose pattern but mix verb styles: sometimes verb-first (get_, ask_, list_, search_, retrieve_, expand_, inspect_) and sometimes query_knowledge_catalog breaks the pattern by using noun-first ordering. Most use snake_case but the verb prefixes are inconsistent (get vs retrieve vs search vs ask vs query all for similar retrieval actions).
9 tools is a reasonable, well-scoped count for a Telegram archive/recall MCP server. The count feels appropriate given the mix of admin/config tools, search tools, and session/catalog inspection tools. Not overly heavy or thin for the domain.
The server covers search, retrieval, context expansion, chat/scope listing, and research session inspection. However, there are notable gaps: no tool to add/configure sync scopes (only list_sync_scopes), no tool to manage allowed chats (only list), and no evident way to trigger or manage syncing/indexing. The surface is read-oriented and administratively incomplete.