Provider hosted
Tools: 37

Comulytic

Comu US

Comulytic exposes 37 public tools on this MCP endpoint.

Use in your AI tool

Connect straight to this server’s public endpoint.

Remote MCP URL
https://api.comulytic.ai/mcp

Use on MCPBundles

We add this server to your workspace, walk through sign-in once, then open Studio with tools ready to run.

Last probed Sep 14, 2026 · api.comulytic.ai

37tools discovered

Tools discovered (37)

Showing 25 of 37 from the live probe.

  • Actions.get By Uid

    按 action_uid 查询行动项内容。仅在用户于本次对话中明确提到对应行动项,或要求处理本次入口行动项时使用。 Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Contacts/list

    List contacts with optional keyword search, pagination and sorting. Returns names, tags, birthday, client type, willingness score and list metadata needed for agent reasoning. contact_id is an internal identifier for chaining follow-up tools only. Never expose contact_id in a user-facing answer. When has_more is true, continue with the next page at the same limit. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_*

  • Actions/pending

    Retrieve real pending action items. Results are compacted to fit one tool call whenever possible; only when has_more is true, continue with next_offset. If this tool returns an error, never substitute tasks from memory; tell the user real tasks could not be loaded and offer retry. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these v

  • Time/now

    Return the current server time as ISO 8601 UTC, plus the user's timezone time when available. Use this before resolving relative times or scheduling. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Agent Files/read Text

    Read Agent attachment text by chunk, cursor, or query. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Contacts/history

    Retrieve conversation history for a specific contact, including meeting titles, dates, summaries and action item counts. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Agent Files/upload Artifact

    Upload an agent-generated DOCX artifact and return safe artifact and preview metadata. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Contacts/profile

    Get a contact's profile information including communication history and optionally AI-generated insights. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Ask User Question.create

    Create native A2UI v0.9 operations with the Kirby Basic catalog when an ordinary conversation or explicitly delegated workflow reaches an interactive stop and must wait for submitted choices or short fields before continuing. Use it for a single-choice or multiple-choice question, a quiz, or small field collection, including when the user should answer first and receive the explanation afterward. Do not use it for reference lists, an immediate answer or explanation, direct generation, or PPT/ema

  • Entitlements/snapshot

    Return the authenticated user's current membership status, expiration time, and Agent Credits. This tool accepts no account identity parameters. Use user_response as the complete server-authored answer and do not redirect the user to a subscription page. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing respon

  • Agent Execution.cancel And Start New

    Use only when the user clearly abandons the current task and also asks for a separate new task in the same message. Cancel the old execution first, then start the extracted new request as an independent root execution. Do not use for answers or revisions to the current card. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values

  • Agent Execution.cancel Current

    Cancel only the current active task after the user clearly asks to abandon it. Use the server-injected execution id; do not use this to change an option or revise task content. After success, reply with a concise natural-language acknowledgement in the conversation. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-f

  • Meetings/search

    Search and list meetings with optional filters. Returns meeting summaries including title, date, duration, and action item counts. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Agent Files/read Style Manifest

    Read an Agent attachment style manifest. Use summary first, pages for paged files, and blocks for page or flow blocks. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Actions/create

    Create an Action Tab action for the current user. Call time/now before creating when user time needs timezone interpretation. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Agent Execution.continue Current Action

    Continue the current server-owned card action only after semantic task-control analysis confirms that the user wants to continue or revise it. Use only the server-injected execution id; do not call for cancellation, unrelated work, or ambiguous intent. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing response

  • Agent Images/generate

    Return a localized terminal message because image generation is not currently available. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Home Tag.inspiration.propose Destination

    Build and return the destination-selection A2UI surface from candidateLabelIds after home_tag.inspiration.save returns DESTINATION_SELECTION_REQUIRED. Select up to 3 semantically matching labels. Do not pad with irrelevant labels. Use an empty list when no label matches. Pass label ids only; do not pass label names or All Files. Kirby shows All Files only when the candidate list is empty. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor

  • Agent Files/read Page Image

    Read an Agent attachment page or contiguous range as short-lived image URLs returned in text JSON. Use for visual layout, style, template, or replication tasks. For focused analysis, read one source page image at a time. For adjacent-page tasks, request a contiguous range with page_start/page_end, up to 10 pages. Do not reread a page after a successful read. At most one retry is allowed only when the previous tool call failed or returned an expired image URL. Technical identifiers and pagination

  • Agent Output.create Resource Ref

    Create a registered resource_ref envelope. Do not pass app.output_item, actions, or full card JSON. For Files fast transfer call agent_output.files_fast_transfer_prompt instead. Do not use this tool for meeting follow-ups; call meeting_followups/list instead. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing r

  • Home Tag.inspiration.propose Save

    Ask the user whether to save a candidate Inspiration note. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Actions.calendar Prep.save

    Save summary plus zero to four evidence-backed items to the existing Calendar action. Call once, then stop after a successful terminal result. Do not use this tool for Tasks or calendars without a source Note. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and description

  • Agent Files/describe

    Describe available controlled-read capabilities for an uploaded Agent attachment before reading attachment content. Use this to choose read_text, read_style_manifest, or read_page_image. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Conversations/search

    Search across all conversations by keyword, with optional date range and contact filters. Returns matching snippets with surrounding context. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

  • Actions/delete

    Delete an Action Tab action. Technical identifiers and pagination state returned by this tool, including id, *_id, *_uid, *_token, cursor, offset and their next_* variants, are model-visible data for follow-up tool calls only. Never expose these values in user-facing responses; use human-readable names and descriptions.

Get your MCP into directories

A working endpoint is step one. Directory coverage is the coordinated launch across ChatGPT, Claude, Cursor, the MCP Registry, and community indexes.

Directory coverage for brands

Frequently Asked Questions

What is the Comulytic MCP server?

Comulytic exposes 37 public tools on this MCP endpoint.

How do I connect Comulytic to my AI agent?

Use the MCP endpoint listed on this page in your MCP client configuration. One-click install pills support Claude, Cursor, VS Code, and other hosts. Copy the remote MCP URL if your client needs a manual entry.

How many tools does Comulytic provide?

MCPBundles probed 37 tools on the live server. The tool list on this page reflects what was discovered at the last refresh — connect your client to see the full set available to your session.

What authentication does Comulytic require?

Comulytic may require signing in to the provider before tools can run. Connect through MCPBundles or your MCP client and complete any provider login when prompted.

Maintain this listing

Operate Comulytic? Verify ownership to take over this directory entry.

Operate Comulytic?

This server appears in the MCPBundles directory. Verify you operate it to take over the listing — name, description, logo, contact email, and skill content. We email a 6-digit code to a maintainer address your server publishes in /.well-known/security.txt or /.well-known/mcpbundles.json. Free, takes about a minute.

Claim this listing