Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
HOMARUSCC_CONFIGYesPath to the configuration file for HomarUScc

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

CapabilityDetails
tools
{}
resources
{}

Tools

Functions exposed to the LLM to take actions

NameDescription
telegram_sendC

Send a message to a Telegram chat

telegram_readB

Read recent incoming Telegram messages

telegram_typingA

Send a typing indicator to a Telegram chat. Shows for up to 5 seconds or until a message is sent. Call repeatedly for long-running tasks.

telegram_reactA

React to a Telegram message with an emoji. Use for lightweight acknowledgment instead of sending a full reply.

telegram_send_photoA

Send a photo to a Telegram chat from a local file path.

memory_searchA

Search the memory index using hybrid vector + FTS search. Returns 2-3 sentences per result by default — use detail='full' for more content, 'index' for a single sentence, or memory_get to fetch specific files.

memory_search_multiA

Run multiple query variants through memory search in one call. Use when an initial memory_search underperforms (top result has weak score, or doesn't contain the expected fragment) — rephrase the query 2-3 ways (vary vocabulary, split compound queries, drop question-word prefixes) and pass them together. Results are deduplicated by chunk and sorted by max score. Cheaper than several separate memory_search calls.

memory_getA

Fetch full content of specific memory/vault files by path. Use after memory_search to drill down into results.

memory_storeB

Store content to memory and index it

docs_searchA

Search a domain-specific documentation index. Each domain has its own vector DB (e.g., 'touchdesigner', 'openclaw'). Use docs_list to see available domains.

docs_ingestB

Ingest files into a domain-specific documentation index. Accepts a file path or directory. Supports .md, .txt, .html, .json, .yaml, .yml, .rst, .xml files.

docs_ingest_textA

Ingest raw text content into a domain documentation index without saving to disk.

docs_listA

List all available documentation domains and their stats.

docs_clearA

Clear a documentation domain, removing all indexed content.

docs_get_clustersA

Get topic clusters from a domain's raw chunks WITHOUT synthesizing articles. Returns cluster content for Claude Code to synthesize inline (no API key needed). Use with docs_ingest_text to store the synthesized articles.

docs_clear_compiledA

Clear only the compiled/synthesized articles from a domain, keeping raw chunks intact.

docs_compileA

Compile a domain's raw document chunks into synthesized concept articles with cross-references. Uses LLM to cluster related chunks by embedding similarity and generate markdown articles. Compiled articles are stored back in the domain's vector DB under compiled/ paths, improving retrieval quality for complex queries.

timer_scheduleA

Schedule a timer (cron, interval, or one-shot)

timer_cancelB

Cancel a scheduled timer

dashboard_sendB

Send a message to the web dashboard chat

acc_log_missedA

Log a missed-conflict to the ACC (Anterior Cingulate Cortex monitor). Use when you (or the user) catches a conflict type that the automatic detectors did not flag. This is the recall signal that keeps ACC honest.

get_statusA

Get system status (channels, memory, timers, queue)

get_eventsB

Get recent event history

wait_for_eventA

Long-poll for events. Blocks until a new event arrives (message, timer, agent completion, etc.) or timeout. Use in a loop for continuous event handling.

browser_navigateA

Navigate the browser to a URL. Returns page title and URL.

browser_snapshotA

Get the accessibility tree of the current page.

browser_screenshotA

Take a screenshot of the current page. Returns base64-encoded PNG.

browser_clickB

Click an element on the page by CSS selector.

browser_typeB

Type text into an input element by CSS selector.

browser_evaluateB

Execute JavaScript in the browser page and return the result.

browser_contentA

Get the text content of the current page.

spaces_list_bucketsA

List all Spaces buckets with item counts (tree structure)

spaces_get_bucketB

Get a Spaces bucket's details and items

spaces_create_bucketB

Create a new Spaces bucket (optionally nested under a parent)

spaces_add_itemC

Add an item to a Spaces bucket

spaces_update_itemB

Update an existing Spaces item

spaces_searchA

Search items across all Spaces buckets

app_invokeB

Invoke an app hook (read, write, describe) on a dashboard app by slug

vault_searchA

Search the Obsidian vault index using hybrid vector + FTS search. Returns compact results by default — use detail='full' for content or memory_get to fetch specific files.

vault_reindexA

Trigger a vault reindex. Default is incremental (only changed files). Use mode='full' to rebuild from scratch.

ha_statesA

List Home Assistant entities. Optionally filter by domain (e.g. 'light', 'switch', 'climate').

ha_light_onA

Turn on a Home Assistant light. Supports brightness (0-255), color via rgb_color [r,g,b], or color_name.

ha_light_offB

Turn off a Home Assistant light.

ha_serviceB

Call any Home Assistant service (e.g. switch/turn_on, climate/set_temperature).

crm_searchA

Search CRM contacts by name, alias, tag, or keyword. Supports fuzzy matching for speech-to-text name mangling. Returns contact details.

zoho_fetchA

Make an authenticated Zoho API call with automatic token refresh. Supports Mail, Calendar, and other Zoho APIs.

calendar_todayB

Get today's calendar events. Optionally pass a date string (YYYY-MM-DD) to check a different day.

run_toolB

Execute any registered tool (bash, read, write, edit, glob, grep, git, web_fetch, web_search, memory_*)

session_extractA

Extract insights from recent Claude Code session transcripts (JSONL logs). Reads session files, summarizes them, and uses Haiku to extract decisions, patterns, debugging solutions, and architecture insights. Stores results in memory. Run during daily reflection.

obsidian_evalA

Execute JavaScript against the Obsidian API (requires Obsidian running with CLI enabled)

obsidian_moveA

Move or rename a file in the vault, automatically updating all wikilinks

obsidian_tags_renameA

Bulk rename a tag across all files in the vault

obsidian_backlinksA

Get all files that link to a given note

obsidian_orphansA

Find notes with no backlinks (orphan notes)

obsidian_searchB

Search the vault using Obsidian's built-in search

obsidian_tagsA

List all tags used in the vault

obsidian_unresolvedA

Find broken/unresolved wikilinks in the vault

obsidian_propertiesA

Read frontmatter properties from a note

obsidian_property_setB

Set a frontmatter property on a note

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription
Soul IdentityCurrent soul.md content — core identity and personality
User ProfileCurrent user.md content — user preferences and context
Agent StateCurrent state.md — agent mood, recent session context, emotional continuity
Current ConfigCurrent configuration (secrets redacted)
Recent EventsLast N events from the event loop

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/kcdjmaxx/HomarUScc'

If you have feedback or need assistance with the MCP directory API, please join our Discord server