vtiger-mcp
MCP server for vtiger CRM that enables AI agents to log timesheets and create/update records in any module via natural language.
README
vtiger-mcp
MCP server + CLI for vtiger CRM. Connect any AI agent (Claude Desktop, Claude Code, Cursor, β¦) to your vtiger CRM to log timesheets and create / update records in any module β in plain English.
It talks to vtiger's built-in REST API (/webservice.php, available in
vtiger 6.x / 7.x). It adds no access rules of its own: every call runs
as the authenticated user, so vtiger enforces that user's role,
profile, and row-level sharing automatically. Each person uses their own
Webservice Access Key, so they only see and change what their account
already allows.
New to the terminal? See SETUP.md for a click-by-click, non-developer guide.
π€ Install with an AI agent (copy-paste prompt)
Don't want to edit config files yourself? Copy the whole prompt below and paste it into an AI agent that can edit files (Claude Code, Cursor, Windsurf, Gemini CLI, OpenAI Codex, VS Code Copilot Agent, β¦). It will ask for your details and set everything up for you.
Install and configure the "vtiger-mcp" MCP server for me (npm package name: vtiger-mcp).
It lets me log vtiger CRM timesheets and create/update records in any module from this agent.
Do this:
1. Verify Node.js 18+ is installed (run: node --version). If it's missing, tell me to install
the LTS from https://nodejs.org and stop.
2. Ask me for these and WAIT for my answers:
- VTIGER_URL: my vtiger base URL, e.g. https://yourcrmdomain.com (no /webservice.php)
- VTIGER_USERNAME: my vtiger login username
- VTIGER_ACCESS_KEY: my "Webservice Access Key" from vtiger -> My Preferences (NOT my password)
3. Detect which MCP client I use and find its config file:
- Claude Desktop: %APPDATA%\Claude\claude_desktop_config.json (Windows) or
~/Library/Application Support/Claude/claude_desktop_config.json (macOS)
- Cursor: ~/.cursor/mcp.json ; Windsurf: ~/.codeium/windsurf/mcp_config.json ;
Gemini CLI: ~/.gemini/settings.json (all use the "mcpServers" JSON key)
- VS Code: ./.vscode/mcp.json (uses a top-level "servers" key with "type": "stdio")
- OpenAI Codex CLI: ~/.codex/config.toml (TOML table [mcp_servers.vtiger])
If you can't tell, ask me which client I'm using.
4. Back up that file, then MERGE IN (do not overwrite my other servers) a server named "vtiger":
command: npx
args: ["-y", "vtiger-mcp"]
env: { VTIGER_URL, VTIGER_USERNAME, VTIGER_ACCESS_KEY, VTIGER_ALLOW_DELETE: "false" }
Use the correct shape for my client (mcpServers JSON, VS Code "servers", or Codex TOML).
5. Keep my access key ONLY in this local config. Never commit it, echo it back in full, or share it.
6. (Optional) Verify auth by running, with those env vars set: npx -p vtiger-mcp vtiger whoami
It should print my vtiger userId.
7. Tell me to fully quit and reopen my client, then try: "list my vtiger modules"
or "log 2 hours to <task name> today".
Prefer to do it manually? See CLIENTS.md for every client's exact config.
Features
- π Per-user auth via each user's Webservice Access Key β row-level access preserved, nothing centralized.
- π§© Any module:
create,update,query,describe,retrieve,deleteβ subject to the user's permissions. - β±οΈ Timesheet helpers for the time-tracking module (create/list/update entries, resolve tasks by name).
- π₯οΈ Ships as both an MCP server (for AI agents) and a CLI.
- π¦ Zero changes to your vtiger install. Node 18+, no native deps.
Requirements
- Node.js 18+
- A vtiger CRM (6.x / 7.x) whose
/webservice.phpendpoint is reachable - Your Webservice Access Key β in the CRM: My Preferences β "Webservice Access Key" (a 16-char string; not your login password)
Install
# Global (gives you the `vtiger` CLI and the `vtiger-mcp` server binary)
npm install -g vtiger-mcp
# β¦or run without installing
npx vtiger-mcp # starts the MCP server (stdio)
Or from source:
git clone https://github.com/ViralP17/vtiger-mcp.git
cd vtiger-mcp
npm install
npm run build
Configure
Set three environment variables (via your MCP client config, or a .env
file for the CLI β see .env.example):
| Variable | Example | Notes |
|---|---|---|
VTIGER_URL |
https://yourcrmdomain.com |
Base URL, no /webservice.php |
VTIGER_USERNAME |
your.username |
Your vtiger login |
VTIGER_ACCESS_KEY |
xxxxxxxxxxxxxxxx |
Webservice Access Key |
VTIGER_ALLOW_DELETE |
false |
Set true to allow delete |
Use as an MCP server
Point your AI client at vtiger-mcp and pass your credentials in env.
Via npx (after npm install -g vtiger-mcp or letting npx fetch it):
{
"mcpServers": {
"vtiger": {
"command": "npx",
"args": ["-y", "vtiger-mcp"],
"env": {
"VTIGER_URL": "https://yourcrmdomain.com",
"VTIGER_USERNAME": "your.username",
"VTIGER_ACCESS_KEY": "your-access-key",
"VTIGER_ALLOW_DELETE": "false"
}
}
}
}
From a source checkout (use node + the built server path):
{
"mcpServers": {
"vtiger": {
"command": "node",
"args": ["/path/to/vtiger-mcp/dist/mcp/server.js"],
"env": { "VTIGER_URL": "https://yourcrmdomain.com", "VTIGER_USERNAME": "your.username", "VTIGER_ACCESS_KEY": "your-access-key" }
}
}
}
- Claude Desktop config:
%APPDATA%\Claude\claude_desktop_config.json(Windows) /~/Library/Application Support/Claude/claude_desktop_config.json(macOS)
π Using Cursor, VS Code, Windsurf, Antigravity, OpenAI Codex CLI, Gemini CLI, or another client? See CLIENTS.md for the exact config file and format for each β it's the same server everywhere.
Then just ask, e.g. "Log 3 hours to the Website Redesign task today as regular hours" or "Show my timesheets this week."
Use as a CLI
export VTIGER_URL=https://yourcrmdomain.com
export VTIGER_USERNAME=your.username
export VTIGER_ACCESS_KEY=your-access-key # or put these in a .env file
vtiger whoami
vtiger modules
vtiger describe TimeControl
vtiger query "SELECT tc_code, date_start, totaltime FROM TimeControl LIMIT 5;"
vtiger log-timesheet --task "Website Redesign" --start 10:00 --end 13:00 --type "Regular Hours" --activity "API work"
vtiger my-timesheets --from 2026-07-01 --to 2026-07-31
vtiger create ProjectTask --set projecttaskname="Write docs" --set projectid=<id>
vtiger update 17x1234 --set description="Revised notes"
Tools exposed to the AI agent
Any module (scoped to your access):
| Tool | What it does |
|---|---|
list_modules |
Modules you can access |
describe_module |
Live fields of a module (names, mandatory, picklists, references) |
query |
Read-only VTQL SELECT |
retrieve_record |
Fetch one record by id |
create_record |
Create a record in any writable module |
update_record |
Partial update of a record |
delete_record |
Delete (only if VTIGER_ALLOW_DELETE=true) |
Timesheet helpers (the time-tracking module):
| Tool | What it does |
|---|---|
whoami |
Your vtiger user id / profile |
list_my_projects |
Projects you can access |
list_my_tasks |
Project tasks (for the related_to link) |
log_timesheet |
Add a daily entry (resolves task, defaults date/owner, computes total) |
list_my_timesheets |
Your entries in a date range |
update_timesheet |
Edit an existing entry |
Notes & tips
- Field names vary per install. vtiger schemas are customizable, so
always
describe_module(orvtiger describe <Module>) first to learn the exact field names before creating/updating. - Ids are vtiger "webservice ids" like
17x1234(<moduleId>x<recordId>). Reference fields (owner,related_to,projectid) expect these ids; the timesheet helpers resolve task names to ids for you. - The timesheet module is configurable (see below). It defaults to
TimeControl; point it at your module without touching code. - Deletes are off by default. Set
VTIGER_ALLOW_DELETE=trueonly if you want the agent/CLI to delete records (and your account has that right). - Security. Your access key is a secret β keep it in your local
.envor client config only. Use HTTPS. Nothing is centralized; each user runs their own copy with their own key.
Configuring the timesheet module (other vtiger installs)
The timesheet helpers default to the TimeControl module with
ProjectTask/Project links. If your install uses different module or
field names, set these env vars (all optional) β no code changes needed:
| Variable | Default | Purpose |
|---|---|---|
VTIGER_TIMESHEET_MODULE |
TimeControl |
Time-entry module |
VTIGER_TASK_MODULE |
ProjectTask |
Module linked from an entry |
VTIGER_PROJECT_MODULE |
Project |
Project module |
Advanced field-name overrides (rarely needed): VTIGER_TS_DATE_START_FIELD,
VTIGER_TS_DATE_END_FIELD, VTIGER_TS_TIME_START_FIELD,
VTIGER_TS_TIME_END_FIELD, VTIGER_TS_RELATED_FIELD, VTIGER_TS_DESC_FIELD,
VTIGER_TS_HOURSTYPE_FIELD, VTIGER_TS_TOTAL_FIELD, VTIGER_TS_OWNER_FIELD,
VTIGER_TASK_NAME_FIELD, VTIGER_TASK_PROJECT_FIELD,
VTIGER_PROJECT_NAME_FIELD. See .env.example for the full
list and defaults.
Whatever your modules are named, the generic
create_record,update_record,query, anddescribe_moduletools always work.
License
MIT Β© ViralP17
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.