filevine-mcp
Full-coverage MCP server for Filevine legal case management, letting you work with matters, contacts, tasks, notes, documents, and billing in natural language from Claude Desktop.
Users — get the current user, list/get users and their tasks, appointments, recently opened/recent projects, and project access
Projects (matters) — list, get, create, update, archive; read vitals and custom forms; manage collections and collection items (CRUD)
Project sub-resources — project contacts (add/update/remove), project team members (add/update/remove), appointments, notes (pin/unpin), deadlines (CRUD), emails, and invoices
Contacts — list/get/create/update, addresses, email addresses, phone numbers, associated projects, country reference list, and tag removal
Tasks — list (all or per project/user), get, create, update, delete, complete/uncomplete, assign, pin/unpin, and snooze
Notes — list, get, create, update, pin/unpin, comment listing/creation/update, and tag removal
Documents — list, get, create, update, delete, download locator, revisions, lock/unlock, move/copy, batch upload and confirm, batch download, search, project association, tag removal, and recently opened
Billing/invoices — get project invoices, create, update, delete, finalize, approve, mark as sent, and fetch the invoice PDF URL
Auth/setup — OAuth 2.0 client-credentials with automatic token refresh; credentials stored in OS keyring with
.envfallback;us,ca, andcjisregions
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@filevine-mcp List my open projects"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Filevine MCP server
Connect Claude and other MCP clients to Filevine to manage projects, contacts, documents, tasks and billing.
Filevine MCP server is a Model Context Protocol server for Filevine, the legal case management platform. It registers 147 tools that read and write Filevine data. It runs over stdio by default, for desktop clients such as Claude Desktop, and offers an opt-in stateless Streamable HTTP mode that implements MCP specification 2026-07-28. Filevine credentials stay on the machine that runs the server: they come from the setup command and your operating system's keyring, never from the client.
Features
Projects (matters): create, update, archive and look up projects, with vitals, custom forms, collections, deadlines and appointments.
Contacts: create, update and look up contacts with addresses, emails, phones and project associations.
Tasks: create, assign, complete, pin and snooze tasks, by project or user.
Notes: create, update and pin notes, add comments, and create or remove tags.
Documents and folders: create, update, lock, unlock, move, copy and search documents, add revisions, batch upload and download, create share links, and organize documents in folders.
Billing: billing items, invoices, payments, project funds and transactions, and rate schedules.
Teams: manage project teams and organization teams.
Email: log email to projects, list project emails, and work with mailroom items.
Webhooks: subscribe to Filevine events for real-time notifications.
Reference data and reports: project types, document series, classifications and reports.
Related MCP server: smokeball-mcp
Tools
The server registers 147 tools.
21 of them delete, archive or remove records, change billing items, or finalize, approve or pay invoices. They take a confirm argument; without confirm=true they return an error and change nothing, for example "archive_project requires confirm=True. Confirm this action with the user before proceeding."
add_contact_to_projectadd_document_revisionadd_document_to_projectadd_email_to_projectadd_team_member_to_projectapprove_invoicearchive_projectassign_taskbatch_download_documentsbatch_upload_documentscomplete_taskconfirm_batch_uploadcopy_documentscreate_billing_itemcreate_collection_itemcreate_contactcreate_documentcreate_foldercreate_hashtagcreate_invoicecreate_mailroom_itemcreate_notecreate_note_commentcreate_paymentcreate_payment_and_applycreate_projectcreate_project_appointmentcreate_project_deadlinecreate_share_linkcreate_taskcreate_teamcreate_webhook_subscriptiondelete_billing_itemdelete_collection_itemdelete_documentdelete_folderdelete_invoicedelete_project_deadlinedelete_share_linkdelete_taskdelete_teamdelete_webhook_subscriptionfinalize_invoiceget_billing_itemget_collection_itemget_contactget_contact_addressesget_contact_emailsget_contact_phonesget_contact_projectsget_countriesget_current_userget_documentget_document_download_locatorget_document_seriesget_folderget_invoice_pdfget_noteget_note_commentget_org_billing_codesget_org_billing_settingsget_org_rate_schedulesget_payment_linkget_projectget_project_billing_codesget_project_billing_settingsget_project_billing_vitalsget_project_contactsget_project_deadlineget_project_formget_project_fund_transactionsget_project_fundsget_project_invoicesget_project_teamget_project_team_memberget_project_transactionsget_project_typeget_project_vitalsget_reportget_share_linkget_taskget_teamget_userget_user_appointmentsget_user_project_accessget_user_recent_projectsget_user_tasksget_webhook_subscriptionincomplete_tasklist_classificationslist_collection_itemslist_contactslist_document_serieslist_documentslist_folderslist_mailroomlist_note_commentslist_noteslist_project_appointmentslist_project_deadlineslist_project_emailslist_project_noteslist_project_taskslist_project_teamslist_project_typeslist_projectslist_recently_opened_documentslist_reportslist_share_linkslist_taskslist_teamslist_userslist_webhook_eventslist_webhook_subscriptionslock_documentmark_invoice_sentmove_documentspin_notepin_note_to_projectpin_taskremove_contact_from_projectremove_project_team_memberremove_tag_from_contactsremove_tag_from_documentsremove_tag_from_notessearch_documentsset_project_rate_schedulesnooze_taskunlock_documentunpin_noteunpin_note_from_projectunpin_taskupdate_billing_itemupdate_collection_itemupdate_contactupdate_documentupdate_folderupdate_invoiceupdate_noteupdate_note_commentupdate_projectupdate_project_contactupdate_project_deadlineupdate_project_formupdate_project_team_memberupdate_taskupdate_webhook_subscription
Prompts and resources
The server also registers three prompts and three resources.
Prompt | What it does |
| Guides a billing workflow for one project: review items, finalize the invoice, record the payment. Takes |
| Morning briefing: open projects needing attention, overdue tasks and pending billing. |
| Reviews a new project's contacts, tasks, documents and billing setup. Takes |
Resource | What it provides |
| All document classification categories configured in your Filevine organisation. |
| All project (matter) types configured in your Filevine organisation. |
| Security notes for the server. |
Requirements
Python 3.10+.
Filevine API credentials from your Filevine organisation administrator or developer portal: a client ID, a client secret and a personal access token (PAT), plus your organization ID.
Your Filevine region:
us,caorcjis.An MCP client such as Claude Desktop.
Installation
Install uv, then clone the repository and install its locked dependencies:
git clone https://github.com/RosenAdvertising/filevine-mcp.git
cd filevine-mcp
uv sync --lockedReleases are also published to PyPI: pip install filevine-mcp installs version 0.3.0, which predates the HTTP mode described below. Install from source to use HTTP mode.
Configuration
Run the setup command once. It prompts for your region, client ID, client secret, personal access token and organization ID, tests the credentials against Filevine, and saves them through the credential store (see Credential storage):
uv run filevine-mcp-setupCheck the connection:
uv run filevine-mcp-verifyServer messages that say to run filevine-mcp-setup mean uv run filevine-mcp-setup from your clone.
The server reads these variables, which you can also set in its environment:
Variable | Required | Default | Purpose |
| Yes (saved by setup) | Credential store | Filevine client ID. |
| Yes (saved by setup) | Credential store | Filevine client secret. |
| Yes (saved by setup) | Credential store | Filevine personal access token. |
| Yes (saved by setup) | Credential store | Filevine organization ID. |
| No |
| Filevine region: |
| No |
| Set to |
| No | Unset | Administrator allowlist for webhook destination hosts (see Webhook destination allowlist). |
Usage with Claude Desktop
Add the server to Claude Desktop's configuration file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS, %APPDATA%\Claude\claude_desktop_config.json on Windows):
{
"mcpServers": {
"filevine": {
"command": "uv",
"args": ["run", "--locked", "--directory", "/absolute/path/to/filevine-mcp", "filevine-mcp"]
}
}
}Replace /absolute/path/to/filevine-mcp with the path of your clone, then restart Claude Desktop. Any other stdio MCP client uses the same command and arguments.
HTTP mode
Stdio is the default. Set FILEVINE_MCP_TRANSPORT=streamable-http to serve the stateless Streamable HTTP transport from MCP specification 2026-07-28 at /mcp. Each request stands alone: no initialization handshake and no Mcp-Session-Id. Clients on earlier protocol versions are served on the same endpoint.
Security: this endpoint has no authentication and no TLS. Anyone who can reach the port can run every tool, including write and delete tools, with this server's vendor credentials. Keep the default loopback bind (
127.0.0.1), or put the server behind an authenticating TLS proxy on a private network.FILEVINE_MCP_ALLOWED_HOSTSandFILEVINE_MCP_ALLOWED_ORIGINSprotect against browser DNS rebinding, not against direct callers. A proxy in front of it needs connection and idle timeouts: a legacy-styleGET /mcpwithAccept: text/event-streamholds a stream open until the client disconnects.
Variable | Default | Purpose |
|
|
|
|
| Bind address. |
|
| Port; must be an integer. |
| unset | Comma-separated |
| unset | Comma-separated |
Filevine credentials come from the same configuration as stdio (see Configuration), never from the request.
FILEVINE_MCP_TRANSPORT=streamable-http PORT=8080 uv run --locked filevine-mcpPoint the MCP client at http://127.0.0.1:8080/mcp. Responses use the SDK's default streaming mode, so a client that disconnects cancels its request.
Error handling
A failed tool call returns an MCP error result (isError) with a fixed message. Errors raised while a tool runs start with Error executing tool <name>: followed by the message in the table below; argument-validation messages are returned without that prefix. The server never passes a Filevine response body, a request URL, a credential or an input value back to the client.
Situation | What the tool returns |
Setup missing or incomplete | "FILEVINE_CLIENT_ID and FILEVINE_CLIENT_SECRET are required. Run: filevine-mcp-setup", "FILEVINE_PAT is required. Run: filevine-mcp-setup" or "FILEVINE_ORG_ID is required. Run filevine-mcp-setup." |
Authorization failed (HTTP 401, after one automatic token request and retry) | "Filevine authorization failed (401). Re-run filevine-mcp-setup to reauthorize." |
Access denied (HTTP 403) | "Filevine access denied: the connected account lacks permission for this action (or the authorization expired; re-run filevine-mcp-setup if so)." |
Rate limited (HTTP 429) | "Filevine rate limit reached. Retry after N seconds." where N comes from the |
Any other HTTP error status | "Filevine request failed with HTTP 500. Check account permissions and request details." (the status varies). |
Timeout or connection failure on a read | "Filevine request timed out or lost its connection. Check connectivity and retry." |
Timeout or connection failure on a write | "Filevine request timed out or lost its connection; the outcome is unknown. Check whether it completed before retrying." |
Response that is not valid JSON | "Filevine returned an unreadable response." |
Missing | "archive_project requires confirm=True. Confirm this action with the user before proceeding." (the tool name varies). |
Webhook destination not approved | "Destination must be a public HTTPS URL approved by FILEVINE_ALLOWED_DESTINATION_HOSTS; configure trusted hostnames." |
Invalid arguments | A message such as "Invalid argument 'limit'; expected integer greater than or equal to 1 and less than or equal to 200." |
Anything else | "Filevine request failed unexpectedly. Check configuration and try again." |
Every Filevine request has a 30-second timeout. On HTTP 429 the server waits for the Retry-After interval (10 seconds when the header is missing) and retries, up to 3 times per request and 60 seconds of total waiting; if the next wait would exceed what is left, the tool returns the rate-limit message at once. On HTTP 401 it requests a new access token once and repeats the request. It does not retry timeouts, connection failures or 5xx responses. A failed resource read returns "Filevine resource could not be read. Check configuration and try again."
At startup the server exits with a message on stderr and a non-zero status when FILEVINE_MCP_TRANSPORT is neither stdio nor streamable-http, when PORT is not an integer, or when a non-loopback FILEVINE_MCP_HOST is set without FILEVINE_MCP_ALLOWED_HOSTS.
Credential storage
By default credentials are stored in your operating system's native secret store
via the cross-platform keyring library:
OS | Backend |
macOS | Keychain |
Windows | Credential Manager |
Linux | Secret Service (GNOME Keyring / KWallet) |
Secrets saved to the keyring use the service name filevine-mcp.
The Filevine access token is cached separately in ~/.filevine-mcp/tokens.json with 0600 permissions, whichever backend holds the credentials.
File fallback. On a host with no keyring backend (e.g. a headless Linux box
without Secret Service), or if you set FILEVINE_MCP_USE_KEYRING=0, credentials
fall back to a ~/.filevine-mcp/.env file with 0600 permissions.
On Windows, the file is stored in the user's profile and protected by Windows'
default per-user access rules. On POSIX, files are created with 0600 permissions
and writes fail closed if private permissions cannot be established.
Read order. A credential already set in the process environment wins. Otherwise the server reads the OS keyring, then the .env file fallback. To make a secret rotated in the keyring take effect, unset FILEVINE_CLIENT_ID and FILEVINE_CLIENT_SECRET in the shell that starts the server.
Authentication notes
Filevine authorizes with a personal access token (PAT) together with your client ID and client secret: the server posts them to your region's identity host (/connect/token) and receives an access token. No browser authorization is required. Access tokens are cached and requested again when they expire. Three regions are supported with separate API and identity hosts:
Region | API Host | Identity Host |
us | api.filevineapp.com | identity.filevine.com |
ca | api.filevineapp.ca | identity.filevine.ca |
cjis | api.filevinegov.com | identity.filevinegov.com |
FILEVINE_REGION and setup accept only us, ca, or cjis. Values are trimmed and lowercased; unknown values fail with an error; they never select a different region.
Example usage in Claude
"List my open projects"
"Create a task on project 456 — send retainer agreement to client"
"Get the billing vitals for project 789"
"Add a note to project 123 — client called re: mediation date"
"Search documents for 'deposition transcript'"
"List all webhook event types available"
Webhook destination allowlist
Set FILEVINE_ALLOWED_DESTINATION_HOSTS in the server environment, for example FILEVINE_ALLOWED_DESTINATION_HOSTS=hooks.firm.example,.integrations.firm.example. Comma-separated exact hosts allow only that host; a leading dot allows the domain and its subdomains. Matching ignores case and trailing dots and normalizes IDNA. An empty or unset list refuses destination URLs before any request. HTTPS, no userinfo, and public literal addresses remain required. This administrator-owned list prevents model-supplied destinations from sending data to arbitrary hosts, including private-address DNS aliases and unapproved redirectors. Approve only hosts whose DNS and redirects the firm trusts; the vendor executes requests later. Tools cannot change this setting.
Webhook destinations in fields_json use an explicit set of destination aliases, validated by the same allowlist even inside nested objects and arrays. Unrelated fields such as securityLevel and jurisdiction pass unchanged.
Testing
The test suite runs offline and needs no Filevine account: every Filevine API call is answered by an in-process fake session. It covers list limits and paging, path-identifier validation, destination URL and region checks, safe error messages and log privacy, credential file handling, the setup and verify commands, the stdio server, and the Streamable HTTP transport including the 2026-07-28 wire format, Host and Origin checks and stateless requests.
uv sync --locked
uv run --locked pytest -qCI runs the suite on every push and pull request to main.
License
MIT. See LICENSE.
Available Tools
147 toolsadd_contact_to_projectB
Add a contact to a project with an optional role.
Inputs: project_id (required), contact_id (required), role, fields_json. Returns: result
add_document_revisionC
Add a new revision to a document.
Inputs: document_id (required), fields_json. Returns: result
add_document_to_projectC
Associate a document with a project.
Inputs: project_id (required), document_id (required). Returns: result
add_email_to_projectC
Add an email record to a project.
Inputs: project_id (required), subject, body, fields_json. Returns: result
add_team_member_to_projectC
Add a team member to a project. fields_json is a JSON object with userId, role, etc.
Inputs: project_id (required), fields_json (required). Returns: result
approve_invoiceC
Approve an invoice.
Inputs: invoice_id (required). Returns: result
archive_projectB
Archive (soft-delete) a project.
Inputs: project_id (required). Returns: result
assign_taskC
Assign a task to a user.
Inputs: task_id (required), assignee_id (required). Returns: result
batch_download_documentsA
Request a batch download of documents. document_ids_csv is comma-separated document IDs.
Inputs: document_ids_csv (required). Returns: result
batch_upload_documentsC
Initiate a batch document upload. fields_json is a JSON object with upload details.
Inputs: fields_json (required). Returns: result
complete_taskC
Mark a task as complete.
Inputs: task_id (required). Returns: result
confirm_batch_uploadC
Confirm a batch document upload. fields_json is a JSON object.
Inputs: fields_json (required). Returns: result
copy_documentsC
Copy documents to a folder. document_ids_csv is a comma-separated list of document IDs.
Inputs: document_ids_csv (required), folder_id (required). Returns: result
create_billing_itemB
Create a billing item (time entry or expense) on a project.
Inputs: project_id (required), description, amount, rate, +2 more. Returns: result
create_collection_itemB
Create an item in a custom collection section. fields_json is a JSON object.
Inputs: project_id (required), selector (required), fields_json (required). Returns: result
create_contactC
Create a new contact.
Inputs: first_name, last_name, email, phone, +2 more. Returns: result
create_documentC
Create a new document record.
Inputs: name, folder_id, fields_json. Returns: result
create_folderC
Create a new document folder.
Inputs: name, parent_folder_id, fields_json. Returns: result
create_hashtagB
Create a hashtag for use in notes and tasks. hashtag is the tag name (without #).
Inputs: hashtag (required), fields_json. Returns: result
create_invoiceC
Create an invoice for a project. fields_json is a JSON object.
Inputs: project_id (required), fields_json. Returns: result
create_mailroom_itemC
Add an item to the mailroom. fields_json is a JSON object.
Inputs: fields_json (required). Returns: result
create_noteC
Create a new note.
Inputs: body, project_id, fields_json. Returns: result
create_note_commentC
Add a comment to a note.
Inputs: note_id (required), body, fields_json. Returns: result
create_paymentC
Record a payment on a project.
Inputs: project_id (required), amount, date, description, +1 more. Returns: result
create_payment_and_applyC
Record a payment and apply it to an invoice.
Inputs: project_id (required), amount, invoice_id, fields_json. Returns: result
create_projectB
Create a new project. Use fields_json for additional fields as a JSON object.
Inputs: project_type_id, name, description, fields_json. Returns: result
create_project_appointmentC
Create an appointment on a project.
Inputs: project_id (required), title, start_date, end_date, +1 more. Returns: result
create_project_deadlineC
Create a deadline on a project.
Inputs: project_id (required), name, due_date, fields_json. Returns: result
create_taskD
Create a new task.
Inputs: body, due_date, project_id, assignee_id, +1 more. Returns: result
create_teamC
Create a new team.
Inputs: name, fields_json. Returns: result
create_webhook_subscriptionB
Create a webhook subscription for a Filevine event.
Inputs: event_name (required), target_url (required), fields_json. Returns: result
delete_billing_itemB
Delete a billing item.
Inputs: billing_item_id (required). Returns: result
delete_collection_itemC
Delete an item from a custom collection section.
Inputs: project_id (required), selector (required), unique_id (required). Returns: result
delete_documentD
Delete a document.
Inputs: document_id (required). Returns: result
delete_folderC
Delete a document folder.
Inputs: folder_id (required). Returns: result
delete_invoiceB
Delete an invoice from a project.
Inputs: project_id (required), invoice_id (required). Returns: result
delete_project_deadlineD
Delete a deadline from a project.
Inputs: project_id (required), deadline_id (required). Returns: result
delete_taskC
Delete a task.
Inputs: task_id (required). Returns: result
delete_teamC
Delete a team.
Inputs: team_id (required). Returns: result
delete_webhook_subscriptionB
Delete a webhook subscription.
Inputs: subscription_id (required). Returns: result
finalize_invoiceC
Finalize an invoice on a project.
Inputs: project_id (required), invoice_id (required). Returns: result
get_billing_itemB
Get a specific billing item on a project.
Inputs: project_id (required), billing_item_id (required). Returns: result
get_collection_itemC
Get a single item from a custom collection section.
Inputs: project_id (required), selector (required), unique_id (required). Returns: result
get_contactC
Get a contact by ID.
Inputs: contact_id (required). Returns: result
get_contact_addressesC
Get all addresses for a contact.
Inputs: contact_id (required). Returns: result
get_contact_emailsB
Get all email addresses for a contact.
Inputs: contact_id (required). Returns: result
get_contact_phonesB
Get all phone numbers for a contact.
Inputs: contact_id (required). Returns: result
get_contact_projectsB
Get all projects associated with a contact.
Inputs: contact_id (required). Returns: result
get_countriesA
Get the list of available countries for contact addresses.
Inputs: no named parameters. Returns: result
get_current_userA
Get the currently authenticated user (me).
Inputs: no named parameters. Returns: result
get_documentC
Get a document by ID.
Inputs: document_id (required). Returns: result
get_document_download_locatorB
Get the download URL/locator for a document.
Inputs: document_id (required). Returns: result
get_document_seriesC
Get a document series template by ID.
Inputs: series_id (required). Returns: result
get_folderA
Get a folder by ID.
Inputs: folder_id (required). Returns: result
get_invoice_pdfC
Get the PDF download URL for an invoice.
Inputs: invoice_id (required). Returns: result
get_noteB
Get a note by ID.
Inputs: note_id (required). Returns: result
get_note_commentB
Get a specific comment on a note.
Inputs: note_id (required), comment_id (required). Returns: result
get_org_billing_codesA
Get all available billing codes for the organisation.
Inputs: no named parameters. Returns: result
get_org_billing_settingsB
Get the organisation's billing settings.
Inputs: no named parameters. Returns: result
get_org_rate_schedulesB
Get all rate schedules for the organisation.
Inputs: no named parameters. Returns: result
get_payment_linkB
Get the online payment link for a project.
Inputs: project_id (required). Returns: result
get_projectC
Get a project by ID.
Inputs: project_id (required). Returns: result
get_project_billing_codesC
Get available billing codes for a project.
Inputs: project_id (required). Returns: result
get_project_billing_settingsB
Get billing settings for a project.
Inputs: project_id (required). Returns: result
get_project_billing_vitalsB
Get billing vitals (summary) for a project.
Inputs: project_id (required). Returns: result
get_project_contactsC
Get all contacts associated with a project.
Inputs: project_id (required). Returns: result
get_project_deadlineB
Get a specific deadline on a project.
Inputs: project_id (required), deadline_id (required). Returns: result
get_project_formB
Get a custom form section on a project by selector name.
Inputs: project_id (required), selector (required). Returns: result
get_project_fundsB
Get trust/retainer fund balances for a project.
Inputs: project_id (required). Returns: result
get_project_fund_transactionsC
Get trust/retainer fund transaction history for a project.
Inputs: project_id (required). Returns: result
get_project_invoicesC
Get all invoices for a project.
Inputs: project_id (required). Returns: result
get_project_teamC
Get the team members assigned to a project.
Inputs: project_id (required). Returns: result
get_project_team_memberB
Get a specific team member on a project.
Inputs: project_id (required), user_id (required). Returns: result
get_project_transactionsB
Get all billing transactions for a project.
Inputs: project_id (required). Returns: result
get_project_typeC
Get a specific project type by ID.
Inputs: project_type_id (required). Returns: result
get_project_vitalsC
Get the vitals (key fields) for a project.
Inputs: project_id (required). Returns: result
get_reportC
Get a specific report by ID.
Inputs: report_id (required). Returns: result
get_taskC
Get a task by ID.
Inputs: task_id (required). Returns: result
get_teamB
Get a team by ID.
Inputs: team_id (required). Returns: result
get_userC
Get a user by ID.
Inputs: user_id (required). Returns: result
get_user_appointmentsC
Get appointments for a user.
Inputs: user_id (required). Returns: result
get_user_project_accessC
Get the list of projects a user has access to.
Inputs: user_id (required). Returns: result
get_user_recent_projectsC
Get recently accessed projects for a user.
Inputs: user_id (required). Returns: result
get_user_tasksC
Get tasks assigned to a user.
Inputs: user_id (required), page, page_size. Returns: result
get_webhook_subscriptionC
Get a specific webhook subscription.
Inputs: subscription_id (required). Returns: result
list_classificationsA
List all document classification categories.
Inputs: no named parameters. Returns: result
list_collection_itemsC
List items in a custom collection section of a project.
Inputs: project_id (required), selector (required), page, page_size. Returns: result
list_contactsC
List all contacts.
Inputs: page, page_size. Returns: result
list_documentsC
List all documents.
Inputs: page, page_size. Returns: result
list_document_seriesA
List all document series templates.
Inputs: no named parameters. Returns: result
list_foldersC
List all document folders.
Inputs: page, page_size. Returns: result
list_mailroomB
List all items in the mailroom.
Inputs: no named parameters. Returns: result
list_note_commentsC
List all comments on a note.
Inputs: note_id (required). Returns: result
list_notesC
List all notes across all projects.
Inputs: page, page_size. Returns: result
list_project_appointmentsB
List all appointments for a project.
Inputs: project_id (required). Returns: result
list_project_deadlinesC
List all deadlines for a project.
Inputs: project_id (required). Returns: result
list_project_emailsC
List emails associated with a project.
Inputs: project_id (required). Returns: result
list_project_notesB
List all notes on a project.
Inputs: project_id (required). Returns: result
list_projectsC
List all projects (matters).
Inputs: page, page_size. Returns: result
list_project_tasksB
List all tasks on a specific project.
Inputs: project_id (required). Returns: result
list_project_teamsA
List all teams assigned to a project.
Inputs: project_id (required). Returns: result
list_project_typesA
List all project (matter) types configured in the organisation.
Inputs: no named parameters. Returns: result
list_recently_opened_documentsB
List recently opened documents for the current user.
Inputs: no named parameters. Returns: result
list_reportsC
List all available reports.
Inputs: no named parameters. Returns: result
list_tasksC
List all tasks across all projects.
Inputs: page, page_size. Returns: result
list_teamsA
List all teams in the organisation.
Inputs: no named parameters. Returns: result
list_usersA
List all users in the organisation.
Inputs: no named parameters. Returns: result
list_webhook_eventsB
List all available webhook event types.
Inputs: no named parameters. Returns: result
list_webhook_subscriptionsA
List all active webhook subscriptions.
Inputs: no named parameters. Returns: result
lock_documentC
Lock a document to prevent editing.
Inputs: document_id (required). Returns: result
mark_invoice_sentC
Mark an invoice as sent.
Inputs: invoice_id (required). Returns: result
move_documentsC
Move documents to a folder. document_ids_csv is a comma-separated list of document IDs.
Inputs: document_ids_csv (required), folder_id (required). Returns: result
pin_noteC
Pin a note.
Inputs: note_id (required). Returns: result
pin_note_to_projectC
Pin a note to a project.
Inputs: project_id (required), note_id (required). Returns: result
pin_taskC
Pin a task.
Inputs: task_id (required). Returns: result
remove_contact_from_projectC
Remove a contact from a project.
Inputs: project_id (required), project_contact_id (required). Returns: result
remove_project_team_memberC
Remove a team member from a project.
Inputs: project_id (required), user_id (required). Returns: result
remove_tag_from_contactsC
Remove a tag from all contacts that have it.
Inputs: tag_name (required). Returns: result
remove_tag_from_documentsA
Remove a tag from all documents that have it.
Inputs: tag_name (required). Returns: result
remove_tag_from_notesB
Remove a tag from all notes that have it.
Inputs: tag_name (required). Returns: result
search_documentsC
Search documents by query string.
Inputs: query (required), page, page_size. Returns: result
set_project_rate_scheduleC
Set the rate schedule for a project.
Inputs: project_id (required), rate_schedule_id (required). Returns: result
snooze_taskC
Snooze a task until a future date (ISO 8601 format).
Inputs: task_id (required), due_date (required). Returns: result
uncomplete_taskC
Mark a completed task as incomplete.
Inputs: task_id (required). Returns: result
unlock_documentA
Unlock a document to allow editing.
Inputs: document_id (required). Returns: result
unpin_noteD
Unpin a note.
Inputs: note_id (required). Returns: result
unpin_note_from_projectC
Unpin a note from a project.
Inputs: project_id (required), note_id (required). Returns: result
unpin_taskC
Unpin a task.
Inputs: task_id (required). Returns: result
update_billing_itemC
Update a billing item on a project.
Inputs: project_id (required), billing_item_id (required), description, amount, +1 more. Returns: result
update_collection_itemC
Update an item in a custom collection section. fields_json is a JSON object.
Inputs: project_id (required), selector (required), unique_id (required), fields_json (required). Returns: result
update_contactC
Update a contact.
Inputs: contact_id (required), first_name, last_name, email, +3 more. Returns: result
update_documentC
Update a document's metadata.
Inputs: document_id (required), name, folder_id, fields_json. Returns: result
update_folderC
Update a document folder.
Inputs: folder_id (required), name, fields_json. Returns: result
update_invoiceC
Update an invoice on a project.
Inputs: project_id (required), invoice_id (required), fields_json. Returns: result
update_noteC
Update a note.
Inputs: note_id (required), body, fields_json. Returns: result
update_note_commentC
Update a comment on a note.
Inputs: note_id (required), comment_id (required), body, fields_json. Returns: result
update_projectC
Update a project. Use fields_json for additional fields as a JSON object.
Inputs: project_id (required), name, description, status, +1 more. Returns: result
update_project_contactC
Update a contact's role or attributes on a project.
Inputs: project_id (required), project_contact_id (required), role, fields_json. Returns: result
update_project_deadlineC
Update a deadline on a project.
Inputs: project_id (required), deadline_id (required), name, due_date, +1 more. Returns: result
update_project_formC
Update a custom form section on a project. fields_json is a JSON object of field values.
Inputs: project_id (required), selector (required), fields_json (required). Returns: result
update_project_team_memberB
Update a team member's role on a project.
Inputs: project_id (required), user_id (required), fields_json. Returns: result
update_taskD
Update a task.
Inputs: task_id (required), body, due_date, fields_json. Returns: result
update_webhook_subscriptionC
Update a webhook subscription.
Inputs: subscription_id (required), target_url, fields_json. Returns: result
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 147 tools
Most tools have clearly distinct purposes due to consistent naming patterns (e.g., get vs list, create vs update). However, the large number of tools increases potential confusion, especially between similar names like list_notes and list_project_notes, or get_user_appointments and get_user_tasks.
Tool names follow a highly consistent verb_noun pattern (e.g., create_project, get_contact, list_documents). Even less common verbs like archive, pin, snooze are still in this pattern. No mixing of styles (e.g., camelCase) is present.
With 147 tools, the server is extremely large. While Filevine is a complex system, this many tools makes it hard for agents to select the right one and increases token usage. Most MCP servers are much smaller; this feels over-scoped.
The tool set covers a wide range of operations for projects, contacts, documents, billing, tasks, notes, and more. Minor gaps exist, such as the absence of delete_contact, delete_note, or create_user, but core workflows are well supported.
Maintenance
Related MCP Connectors
Law firm management MCP: manage cases, clients, tasks, calendar and documents via Claude AI.
MCP server unifying ERPs, CRMs, APIs and knowledge base for Claude, ChatGPT and Gemini.
Clio MCP server for ChatGPT, Claude and any AI agent: search matters, log time entry, create task, add matter note, create calendar entry and list bills in your own Clio account. Cotlet works in the account through your Cloud Browser, a real Chrome where you sign in to Clio yourself; Cotlet does not store your password and you need no Clio API key. About the official Clio API: Yes. Clio has a REST API, the Clio Manage API v4, with OAuth 2.0 (checked 6 Nov 2026). Setup and FAQ: https://cotlet.dev/integrations/clio-api-mcp-server
B2B, GovCon & LegalTech MCP server for live RFPs, contract parsing, compliance & proposal tools.
Related MCP Servers
- AlicenseAqualityCmaintenanceMCP server that connects Clio Manage to Claude via OAuth, enabling law firm management tasks like matter lookup, time tracking, billing, calendar, and document retrieval through natural language.41MIT
- AlicenseCqualityBmaintenanceMCP server providing full Smokeball API coverage for law firm practice management, enabling natural language interactions with matters, contacts, tasks, billing, and more via Claude Desktop.1001MIT
- AlicenseCqualityBmaintenanceMCP server for Bill4Time providing API coverage for legal billing and time tracking, enabling natural language interaction with clients, projects, time entries, invoices, payments, and more from Claude Desktop.591MIT
- AlicenseAqualityCmaintenanceAn open-source MCP server connecting Claude to Filevine practice management platform. Access cases, contacts, documents, notes, tasks, and custom PI data from Claude with automatic rate limiting and audit logging.151MIT