Offers community support through a Discord server for real-time help and discussions
Offers containerized deployment of the ToolFront MCP server through Docker images
Provides connection to DuckDB databases, allowing AI agents to execute queries and explore DuckDB data
Provides issue tracking and repository hosting for the ToolFront project
Enables GitHub Copilot to connect to your databases, providing context about tables, schemas, and query patterns
Enables connections to MySQL databases, allowing AI agents to query and analyze MySQL data
Provides connection to PostgreSQL databases, allowing AI agents to query and work with PostgreSQL data
Enables connections to Snowflake data warehouses, allowing AI agents to query and analyze Snowflake data
Provides connection to SQLite databases, allowing AI agents to query and analyze SQLite data
ToolFront
The missing link between AI and big data
It's hard to ask AI about your data. Out-of-the-box models struggle to understand large databases and APIs, while fine-tuned models are expensive and brittle. ToolFront solves this by helping AI models discover and learn about your data on the fly, so they can quickly answer your questions.
Features
- 🌊 Seamless: Connect AI to all your databases, warehouses, and APIs.
- ⚡ Instant: Get up and running in seconds with a single command.
- 🧩 Pluggable: Works with any LLM, agent framework, and IDE that supports MCP.
- 🧠 Self-improving: Your AI learns from experience, becoming smarter and faster over time.
- 🔒 Secure: Your data stays local, private, and under your control.
You need relevant data to make AI fast and accurate. ToolFront's learning API (🧠) automatically gathers structural information about your databases and APIs to help your AI get better over time. This feature is in beta, and we'd love your feedback!
Quickstart
ToolFront runs on your computer through an MCP server, a secure protocol to connect apps to LLMs.
Prerequisites
- uv or Docker to run the MCP server (we recommend uv)
- URLs of your databases and APIs - see below
- API key (optional) to activate self-improving AI with the learning API
Run inside your AI Framework or IDE
First, create an MCP config by following the instructions for your chosen framework or IDE.
IDE | Setup Instructions | Install with UV | Install with Docker |
---|---|---|---|
Cursor | Settings → Cursor Settings → MCP Tools (or create .cursor/mcp.json file) | 🔗 Quick Install | 🔗 Quick Install |
GitHub Copilot (VSCode) | Copilot icon → Edit preferences → Copilot Chat → MCP | 🔗 Quick Install | 🔗 Quick Install |
Then, add as many database and API URLs to the MCP configuration as you need:
You're all set! You can now ask your AI agents about your data.
Tip
By default, uvx toolfront[all]
installs all package extras. For a lighter setup, you can directly install the extras you need e.g. uvx toolfront[postgres,mysql]
. See Databases for the full list of extras.
Run directly
Spin up the ToolFront MCP server with SSE or stdio using the --transport
flag.
Optionally, activate self-improving AI by passing your learning API key with the --api-key "YOUR-API-KEY"
flag.
Tip
Version control: To pin specific versions of ToolFront, use "toolfront[all]==0.1.x"
for UV or antidmg/toolfront:0.1.x
for Docker.
Data Sources
ToolFront supports both databases and APIs:
Databases
See the table below for the list of supported databases, extras (e.g., uvx "toolfront[snowflake,databricks]"
) and connection URL formats.
Database | Extras | URL Format |
---|---|---|
BigQuery | bigquery | bigquery://{project-id}?credentials_path={path-to-account-credentials.json} |
Databricks | databricks | databricks://token:{token}@{workspace}.cloud.databricks.com/{catalog}?http_path={warehouse-path} |
DuckDB | duckdb | duckdb://{path-to-database.duckdb} |
MySQL | mysql | mysql://{user}:{password}@{host}:{port}/{database} |
PostgreSQL | postgresql , postgres | postgres://{user}:{password}@{hostname}:{port}/{database-name} |
Snowflake | snowflake | snowflake://{user}:{password}@{account}/{database} |
SQL Server | mssql , sqlserver | mssql://{user}:{password}@{server}:{port}/{database} |
SQLite | sqlite | sqlite://{path-to-database.sqlite} |
Don't see your database? Submit an issue or pull request, or let us know in our Discord!
Tip
Working with local data files? Add duckdb://:memory:
to your config to analyze local Parquet, CSV, Excel, and JSON files.
APIs
ToolFronts supports virtually all APIs that have an OpenAPI or Swagger specification. See the table below for a list of common APIs and their specification URLs.
API | Specification URL |
---|---|
Wikipedia | https://en.wikipedia.org/api/rest_v1/?spec |
GitHub | https://raw.githubusercontent.com/github/rest-api-description/main/descriptions/api.github.com/api.github.com.json |
Stripe | https://raw.githubusercontent.com/stripe/openapi/master/openapi/spec3.json |
Slack | https://raw.githubusercontent.com/slackapi/slack-api-specs/master/web-api/slack_web_openapi_v2.json |
Discord | https://raw.githubusercontent.com/discord/discord-api-spec/main/specs/openapi.json |
Note
Authentication: For APIs that require authentication, append your API key or token to the specification URL (e.g., https://api.com/openapi.json?token=YOUR-API-TOKEN
). ToolFront will automatically detect and use the authentication parameters in the appropriate places.
Tools
MCP tools are functions that AI agents can call to interact with external systems. ToolFront comes with ten database tools:
Tool | Description | Requires API Key |
---|---|---|
discover | List all configured databases and APIs | ✗ |
search_endpoints | Search API endpoints by pattern or similarity | ✗ |
search_tables | Search database tables by pattern or similarity | ✗ |
sample_table | Get sample rows from a database table | ✗ |
inspect_table | Show structure and columns of a database table | ✗ |
inspect_endpoint | Show structure and parameters of an API endpoint | ✗ |
query_database | Run read-only SQL queries against databases | ✗ |
request_api | Make requests to API endpoints | ✗ |
search_queries | Retrieve and learn from relevant query samples | ✓ |
search_requests | Retrieve and learn from relevant requests samples | ✓ |
FAQ
ToolFront stands out with multi-database support, self-improving AI, and a local-first architecture.
Multi-database: Instead of being limited to a single database, ToolFront connects all your databases and APIs in one place.
Self-improving: ToolFront learning API helps your AI agents get smarter and faster over time.
Local-first: Cloud solutions compromise your data and rack up egress fees. ToolFront keeps everything local.
The learning API uses in-context learning, a novel training-free learning framework pioneered by OpenAI. By augmenting your LLM's context query and request samples, your agents can reason by analogy over your databases and APIs to quickly arrive at the correct answer.
- Local execution: All database connections and queries run on machine.
- No secrets exposure: Database secrets are never shared with LLMs.
- Read-only operations: Only safe, read-only database queries are allowed.
- No data transmission: Your database content never leaves your environment.
- Secure MCP protocol: Direct communication between agents and databases without third-party storage.
Run the uv run toolfront[all]
or docker run
commands with your database URLs directly from the command line. ToolFront automatically tests all connections when starting and will display detailed errors if a connection fails. If you're still having trouble, double-check your database and API URLs using the examples in the Databases section above.
Support & Community
Need help with ToolFront? We're here to assist:
- Discord: Join our community server for real-time help and discussions
- Issues: Report bugs or request features on GitHub Issues
Contributing
See CONTRIBUTING.md for guidelines on how to contribute to ToolFront.
License
ToolFront is released under the MIT License. This means you are free to use, modify, and distribute the software, subject to the terms and conditions of the MIT License. For the full license text, see the LICENSE file in the repository.
This server cannot be installed
local-only server
The server can only run on the client's local machine because it depends on local resources.
Securely connects AI agents to multiple databases simultaneously while enabling collaborative learning from team query patterns, all while keeping data private by running locally.
Related MCP Servers
- AsecurityFlicenseAqualityEnables AI models to perform MySQL database operations through a standardized interface, supporting secure connections, query execution, and comprehensive schema management.Last updated -73522JavaScript
- -securityFlicense-qualityIntegrates with the AgentCraft framework to enable secure communication and data exchange between AI agents, supporting both premade and custom enterprise AI agents.Last updated -Python
- -securityFlicense-qualityEnables secure interaction with MySQL databases, allowing AI assistants to list tables, read data, and execute SQL queries through a controlled interface.Last updated -Python
- -securityFlicense-qualityA powerful server that enables AI agents to interact with MySQL databases, execute SQL queries, and manage database content through a simple interface.Last updated -254JavaScript