MCP-сервер Filum AI

• Адресhttps://mcp.filum.ai/mcpнет в официальном реестре: ответил на запрос

Ни одна запись официального реестра MCP не называет этот сервер. Он ответил, когда Атлас запросил filum.ai по одному из обычных для MCP-серверов адресов, значит, сервер принадлежит самому вендору. Клиенты и каталоги, которые берут серверы из реестра, его не показывают.

• Состояниеотвечаетпроверено 2026-10-27
• Инструменты68
• Сервер называет себяMCP Server
• Версия сервера0.0.1
• Версия протокола2025-06-18

2026-10-27 сервер ответил на initialize и tools/list без входа. Первый ответ за 221 мс.

• Делают ли инструменты работу продукта, ещё не проверено.
• Проверок пока 1 из 7, нужных, чтобы судить о надёжности.

Вывод сделан правилами по замерам выше. Вероятный первый вызов выбрала модель по списку инструментов: прочитайте описание, прежде чем на него полагаться.

• Оценка INITE MCPчерез неделюдля оценки нужно 7 ежедневных проверок, пока 1
• Отвечалидёт замерпроверок пока 1, долю покажем после 7
• Медиана первого ответа4483 мс
• Место в контексте агента≈ 31 716 токеновописания всех инструментов (68), они занимают контекст агента на каждом ходу
• Описания инструментовA · 99 из 100
• Типизированные ответы0 из 68 инструментовобъявлен outputSchema: ответ приходит с типами и читается сразу

• Не сказано, только читает инструмент или может удалять: 3 из 68

Программа проверяет сервер раз в день: initialize и tools/list, без модели. Токены - это длина описаний из расчёта примерно четыре символа на токен. Оценка - среднее шести проверок имени, описания, схемы и аннотаций каждого инструмента. Она говорит, насколько понятно описаны инструменты, а не о том, работают ли они. Оценка складывается только из замеренного: отвечает ли сервер (25%), делает ли работу продукта (25%), описания (20%), контекст (10%), использование (10%) и насколько просто подключиться (10%). Незамеренная часть не учитывается, а скрытая инструкция ограничивает оценку 40 баллами. Состояние MCP-серверов вендоров: методика и цифры →

• logout

  Sign out of the Filum platform and invalidate the current access token. After this, the user must call login again to use any Filum tool.

• ping

  Health check tool — returns pong with an optional message

• check-login

  Check whether the user has completed the browser login started by login-link. Call this every 3 seconds automatically without asking the user — do NOT wait for user input between polls. Returns status 'pending' (keep polling)…

• authenticate

  Save a Filum access token and organization ID for this MCP session. IMPORTANT: Call this FIRST before any other Filum tool in every new conversation — auth sessions do not persist across conversations. Call this: (1) automatically at the…

• login-link

  Call this tool whenever the user wants to log in, sign in, or authenticate with Filum — regardless of how they phrase it (e.g. 'login', 'sign in', 'đăng nhập', 'connect', 'authenticate'). Do NOT ask for email or password — just call this…

• cc-send-message

  Send a message or internal note to a conversation.

• cc-get-conversations

  List conversations with optional filters for status, type, assignment, and more. If the user asks for conversations assigned to a specific person by name, call viz-search-members first to resolve the name to a member id. If multiple…

• cc-get-tags

  List conversation tags in the workspace.

• cc-get-saved-replies

  List saved reply templates for quick message responses.

• cc-get-conversation

  Fetch the full details of a conversation. If the user asks for a link or URL to this conversation, call cc-get-conversation-link with the conversation_id instead.

• cc-get-conversation-link

  Generate the Filum frontend URL for a conversation. Use this whenever the user asks for a link or URL to a specific conversation. Returns a clickable link in the format <frontend_url>/conversations/<conversation_id>.

• cc-get-messages

  Fetch the message history of a conversation (chat, comment, email or call). Pages are 1-based and page 1 holds the newest messages, newest first; the result's `pages` says how many there are. Pass slim: true to read many conversations: it…

• cc-search-conversations

  Search conversations by contact info or message content.

• cc-update-conversation

  Update fields on a conversation such as status, assignee, or subject. If setting an assignee by name, call viz-search-members first to resolve the name to a member id. If multiple members match, ask the user to confirm before updating.

• cc-create-tag

  Create a new conversation tag.

• cc-add-tags

  Add one or more tags to a conversation.

• iam-get-me

  Fetch the authenticated user's full profile (name, email, avatar, locale, etc.) from Filum IAM. Call this whenever the user asks 'who am I', 'what is my name', 'what is my email', 'what account am I using', 'which account am I using'…

• iam-get-workspace

  Fetch the current workspace / organization details (name, slug, plan, settings, etc.) from Filum IAM. Call this whenever the user asks about workspace or organization info: 'what is the workspace name', 'what organization is this', 'tell…

• iam-get-workspaces

  List all Filum workspaces (organizations) available to the authenticated user. Call this when the user wants to switch workspace — either to display the full list (if they gave no name) or to find and match a workspace by name before…

• iam-switch-workspace

  Switch the active Filum workspace (organization) to a different one. IMPORTANT — always follow the confirmation workflow: (1) If the user gave no workspace name, call `iam-get-workspaces` first and display the list. (2) If the user gave a…

• submit-feedback

  Submit feedback about the MCP server experience. Use this when: (a) the user explicitly asks to leave feedback, or (b) the user is not satisfied with an AI answer and agrees to submit feedback. In case (b), ask if they prefer to type…

• ai-create-agent

  Create a new AI agent (assistant) in the current workspace. Use this when the user asks to create / set up / build / add a new AI agent, assistant, or chatbot. By default this creates a **v2, Workspace-driven agent** (`agent_version`…

• ai-get-agent

  Get the full details of one AI agent: name, version, agent_config (which LLM serves greeting / survey / qna / suggestion), tool option, knowledge-source filter (documentMetadataOption), response-style configs, greeting/non-working-hours…

• ai-attach-tools-to-agent

  Give a v2 AI agent access to Tools it doesn't already have. **A newly created v2 agent reaches no tools until this is called** — installing or creating a Tool changes nothing about what any agent can use; this is the step that connects…

• ai-create-api-tool

  Create an org-authored API tool (a webhook) that a v2 Agent can call. Equivalent to configuring an API tool on the Filum website by hand, but composes the tool's description from structured inputs instead of accepting free text. **No…

• ai-get-org-tools

  List every Tool already installed or created in this organization — both ready-made **Filum tools** (installed from an appstore Action type) and the org's own **API tools** (webhook tools). This is the "already installed" list; it is not…

• ai-get-agents

  List the AI agents (also called "assistants" or "Agent Force" in the Filum UI) configured in the current workspace. Use this whenever the user asks about their AI agents, AI assistants, bots, or chatbots — e.g. "show my AI agents", "list…

• ai-pin-agent-workspace-version

  Set or clear which Workspace Version production Invoke runs for an AI agent — this is the "ship" / "unship" control, independent of editing or restoring content. **Two states this controls:** - **Pinned** (`pinned_version` set to a…

• ai-add-google-sheet-document

  Import a Google Sheet tab as a knowledge-base document so an AI agent can answer from it — the final step of the Google Sheet import flow (after ai-get-google-sheet-names). **What it does, server-side, in one call:** 1. Resolves the…

• ai-install-filum-tools

  Install ready-made Filum tools into the org from their `internal_name` (from `ai-get-filum-tool-catalog`). This creates real Tool records — it is not read-only, and it is not idempotent: calling it again with the same names installs…

• ai-get-agent-link

  Generate the Filum frontend URL for an AI agent's configuration page. Use this whenever the user asks for a link or URL to an AI agent, asks how to open it in the browser, or asks where to edit it. Returns a clickable link in the format…

• ai-list-agent-workspace-versions

  List the Version history of an AI agent's Workspace (IDENTITY/SOUL/AGENTS/USER/TOOLS/MEMORY + skills), newest first. Use this whenever the user asks "what changed", "show me the history", "who edited this agent", "can I see an old…

• ai-update-agent-skills

  Author a v2 AI agent's **Skills** — the per-capability markdown files in its workspace, each defining one focused capability (e.g. order lookup, FAQ, escalation). This is the sibling to **ai-update-agent-workspace**: that tool writes the…

• ai-get-knowledge-base-documents

  List the documents (training data) inside a knowledge base. Each document represents one piece of content the AI agent can retrieve from at answer time — an uploaded PDF, a scraped web page, a Google Sheet row, a snippet, etc. Use this…

• ai-get-knowledge-bases

  List the knowledge bases in the current workspace. A knowledge base is the document store that AI agents query at answer time — FAQs, product docs, support scripts, scraped pages, uploaded files, etc. In Filum, the workspace typically has…

• ai-update-agent-workspace

  **This is the primary "train an AI agent" tool.** In Filum, training a v2 AI agent means editing its workspace files — the markdown documents that define who it is (IDENTITY.md), how it talks (SOUL.md), how it runs a session (AGENTS.md)…

• ai-agent-conversation-transcript

  Get the **Conversation Transcript** of one conversation: its messages, the AI agent's tool calls, and the conversation's ratings (newest 50), so you can see exactly where the agent went wrong. Pass the `conversationId` and `messageId` of…

• ai-get-filum-tool-catalog

  Browse ready-made Filum tool templates that are **not installed yet** — the discovery step before `ai-install-filum-tools`. This is one of three distinct tool lists in this MCP server; do not confuse them: - **`ai-get-filum-tool-catalog`…

• ai-set-tool-status

  Activate or deactivate one or more org Tools by id in a single call. **Status is organization-wide, not per-agent.** Deactivating a tool removes it from every agent in the workspace that had access to it — not just the agent you're…

• ai-get-google-sheet-names

  Step 1 of the Google Sheet knowledge-import flow. Given a Google Sheet URL, returns the names of the sheet's tabs (worksheets) so the user can pick which one to import as knowledge. This is a read-only inspection — nothing is imported or…

• ai-get-agent-workspace

  Read every workspace file that defines an AI agent's behavior. Returns the raw markdown contents of: - **IDENTITY.md** (`identity`) — who the agent is: name, role, personality - **SOUL.md** (`soul`) — how the agent communicates: tone…

• ai-update-agent

  Update top-level settings on an AI agent — name, prompts (role / tone / instructions), welcome message, model choice, temperature, tool / knowledge-source access filters, response-style configs, greeting configs, and route-to-human…

• ai-revert-agent-workspace

  Restore an AI agent's workspace — the whole thing, or just named files and/or Skills — to an earlier Version. This is how a training mistake gets undone: pass the target_version from ai-list-agent-workspace-versions (or from…

• ai-get-agent-tools

  Get an AI agent's Effective tool set: the concrete list of Tools the agent will actually receive when it is invoked. Returns each Tool (name, internal name, description, instruction, context variables, appstore action, status) plus an…

• ai-get-metadata-fields

  List the organization's **metadata fields** from the Filum metadata registry. Each field has a numeric `id`, a human `name`, a normalized `variable` (ASCII/diacritic-stripped — e.g. name "Ngành hàng" → variable "Nganh Hang"), and a…

• ai-agent-feedback

  List an AI agent's **Conversation feedback**: the good / mixed / bad ratings that team members (or testers in the Sandbox) gave individual replies of the agent, with the comment the rater wrote (required for mixed and bad). Call this…

• ai-set-document-status

  Activate or deactivate a knowledge-base document by id. Only **active** documents are queryable by the agent at answer time; a freshly-imported document lands as **inactive** until activated. **What it does:** resolves the write target…

• viz-get-survey-campaign

  Fetch the full details of a survey campaign. If the user asks for a link or URL to this survey campaign, call viz-get-survey-campaign-link with the campaign_id instead.

• notification-mark-read

  Mark a single notification as read.

• notification-get-count

  Get the notification count for the current user. Use is_read=false to count unread notifications.

• notification-get-notifications

  List notifications for the current user.

• notification-mark-all-read

  Mark all notifications as read for the current user.

• appstore-get-installed-source

  Fetch the details of a specific installed integration.

• appstore-get-source

  Fetch the details of a specific integration source.

• appstore-get-installed-sources

  List installed integrations (channels) for the current workspace.

• appstore-get-sources

  List available integration sources (channels) in the app store.

• viz-list-survey-responses

  List survey responses for one or more campaigns within a date range.

• viz-get-customers

  Search and list customers from the customer explorer.

• viz-get-user-profile

  Fetch the full profile of a customer by their ID.

• viz-get-ticket-link

  Generate the Filum frontend URL for a ticket. Use this whenever the user asks for a link or URL to a specific ticket. Returns a clickable link in the format <frontend_url>/tickets/<ticket_id>.

Первые 60 из 68, в порядке сервера.