📝 Add MCP/function call to README.md

Author: julien040

.gitignore

@@ -61,3 +61,4 @@ fabric.properties
 .DS_Store
 
 dist/
+.aider*

README.md

@@ -10,30 +10,37 @@
 [![Dynamic JSON Badge](https://img.shields.io/badge/dynamic/json?url=https%3A%2F%2Fregistry.anyquery.dev%2Fv0%2Fquery%2F&query=%24.queries_count&style=flat&label=Queries%20from%20the%20hub&cacheSeconds=3600&link=https%3A%2F%2Fanyquery.dev%2Fqueries)](https://anyquery.dev/queries)
 [![Go Reference](https://pkg.go.dev/badge/github.com/julien040/[email protected]/namespace.svg)](https://pkg.go.dev/github.com/julien040/anyquery/namespace)
 
-Anyquery is a SQL query engine that allows you to run SQL queries on pretty much anything. It supports querying [JSON](https://anyquery.dev/docs/usage/querying-files/#json), [CSV](https://anyquery.dev/docs/usage/querying-files/#csv), [Parquet](https://anyquery.dev/docs/usage/querying-files/#parquet), SQLite, [Airtable bases](https://anyquery.dev/integrations/airtable/), [Google Sheets](https://anyquery.dev/integrations/google_sheets/), [Notion databases](https://anyquery.dev/integrations/notion/), and [log files](https://anyquery.dev/docs/usage/querying-log/) using [Grok](https://www.elastic.co/guide/en/elasticsearch/reference/current/grok.html), among others. It also supports running SQL queries on [remote files](https://anyquery.dev/docs/usage/querying-files/#remote-files) (HTTP, S3, GCS), local apps ([Apple Notes](https://anyquery.dev/integrations/notes/), [Apple Reminders](https://anyquery.dev/integrations/reminders/), [Google Chrome Tabs](https://anyquery.dev/integrations/chrome/), etc.), and databases ([MySQL](https://anyquery.dev/docs/database/mysql/) and [PostgreSQL](https://anyquery.dev/docs/database/postgresql/)). It's built on top of [SQLite](https://www.sqlite.org) and uses [plugins](https://anyquery.dev/integrations/) to extend its functionality.
+Anyquery is a SQL query engine that allows you to run SQL queries on pretty much anything. It supports querying [files](https://anyquery.dev/docs/usage/querying-files/), [databases](https://anyquery.dev/docs/database), and [apps](https://anyquery.dev/integrations) (e.g. Apple Notes, Notion, Chrome, Todoist, etc.). It's built on top of [SQLite](https://www.sqlite.org) and uses [plugins](https://anyquery.dev/integrations) to extend its functionality.
 
-Moreover, it can act as a [MySQL server](https://anyquery.dev/docs/usage/mysql-server/), allowing you to run SQL queries from your favorite MySQL-compatible client (e.g. [Looker Studio](https://anyquery.dev/connection-guide/looker-studio/), [DBeaver](https://anyquery.dev/connection-guide/dbeaver/), [TablePlus](https://anyquery.dev/connection-guide/tableplus/), [Metabase](https://anyquery.dev/connection-guide/metabase/), etc.).
+It can also connect to [LLMs](https://anyquery.dev/llm) (e.g. ChatGPT, Claude, Cursor, TypingMind, etc.) to allow them to access your data.
+
+Finally, it can act as a [MySQL server](https://anyquery.dev/docs/usage/mysql-server/), allowing you to run SQL queries from your favorite MySQL-compatible client (e.g. [TablePlus](https://anyquery.dev/connection-guide/tableplus/), [Metabase](https://anyquery.dev/connection-guide/metabase/), etc.).
 
 ![Anyquery header](https://anyquery.dev/images/release-header.png)
 
 ## Usage
 
-The [documentation](https://anyquery.dev/docs/usage/running-queries) provides detailed instructions on how to run queries with Anyquery.
-But let's see a quick example. Type `anyquery` in your terminal to open the shell mode. Then, run the following query:
+### Connecting LLM
 
-```sql
--- List all repositories of asg017 related to SQLite
-SELECT full_name, stargazers_count, pushed_at FROM github_repositories_from_user('asg017') WHERE name LIKE '%sqlite%';
+LLMs can connect to Anyquery using the [Model Context Protocol (MCP)](https://anyquery.dev/docs/reference/commands/anyquery_mcp). This protocol provides context for LLMs that support it. You can start the MCP server with the following command:
 
--- Count rows of a remote 75MB CSV file
-SELECT count(*) FROM read_csv('https://raw.githubusercontent.com/datadesk/california-coronavirus-data/master/latimes-place-totals.csv', header=true);
+```bash
+# To be started by the LLM client
+anyquery mcp
+# To connect using an HTTP and SSE tunnel
+anyquery mcp --tunnel
+```
 
--- Insert into a Notion database all repositories of nalgeon related to SQLite
-INSERT INTO notion_database(repo, stars, last_push) SELECT full_name, stargazers_count, pushed_at FROM github_repositories_from_user('nalgeon') WHERE description LIKE '%sqlite%';
+You can also connect to clients that supports function calling (e.g. ChatGPT, TypingMind). Refer to each [connection guide](https://anyquery.dev/integrations#llm) in the documentation for more information.
 
--- Close all tabs of the datasette documentation
-DELETE FROM chrome_tabs WHERE url LIKE '%datasette%';
-```
+![5ire example](https://anyquery.dev/images/docs/llm/5ire-final.png)
+
+### Running SQL queries
+
+The [documentation](https://anyquery.dev/docs/usage/running-queries) provides detailed instructions on how to run queries with Anyquery.
+But let's see a quick example. Type `anyquery` in your terminal to open the shell mode. Then, run the following query:
+
+![Anyquery SQL examples](https://anyquery.dev/images/anyquery_examples.sql.png)
 
 You can also launch the MySQL server with `anyquery server` and connect to it with your favorite MySQL-compatible client.
 

website/public/images/anyquery_examples.sql.png

@@ -59,7 +59,9 @@ anyquery -d ":memory:"
 * [anyquery alias](../anyquery_alias)	 - Manage the aliases
 * [anyquery completion](../anyquery_completion)	 - Generate the autocompletion script for the specified shell
 * [anyquery connection](../anyquery_connection)	 - Manage connections to other databases
+* [anyquery gpt](../anyquery_gpt)	 - Open an HTTP server so that ChatGPT can do function calls
 * [anyquery install](../anyquery_install)	 - Search and install a plugin
+* [anyquery mcp](../anyquery_mcp)	 - Start the Model Context Protocol (MCP) server
 * [anyquery plugins](../anyquery_plugins)	 - Print the plugins installed on the system
 * [anyquery profiles](../anyquery_profiles)	 - Print the profiles installed on the system
 * [anyquery query](../anyquery_query)	 - Run a SQL query

website/public/images/release-header.png

@@ -0,0 +1,36 @@
+---
+title: anyquery gpt
+description: Learn how to use the anyquery gpt command in Anyquery.
+---
+
+Open an HTTP server so that ChatGPT can do function calls
+
+### Synopsis
+
+Open an HTTP server so that ChatGPT can do function calls. By default, it will expose a tunnel to the internet.
+By setting the --host or --port flags, you can disable the tunnel and bind to a specific host and port. In this case, you will need to configure your LLM to connect to this host and port.
+
+```bash
+anyquery gpt [flags]
+```
+
+### Options
+
+```bash
+  -c, --config string       Path to the configuration database
+  -d, --database string     Database to connect to (a path or :memory:)
+      --extension strings   Load one or more extensions by specifying their path. Separate multiple extensions with a comma.
+  -h, --help                help for gpt
+      --host string         Host to bind to. If not empty, the tunnel will be disabled
+      --in-memory           Use an in-memory database
+      --log-file string     Log file
+      --log-format string   Log format (text, json) (default "text")
+      --log-level string    Log level (trace, debug, info, warn, error, off) (default "info")
+      --port int            Port to bind to. If not empty, the tunnel will be disabled
+      --read-only           Open the SQLite database in read-only mode
+      --readonly            Open the SQLite database in read-only mode
+```
+
+### SEE ALSO
+
+* [anyquery](../anyquery)	 - A tool to query any data source

website/src/content/docs/docs/reference/Commands/anyquery.md

@@ -0,0 +1,41 @@
+---
+title: anyquery mcp
+description: Learn how to use the anyquery mcp command in Anyquery.
+---
+
+Start the Model Context Protocol (MCP) server
+
+### Synopsis
+
+Start the Model Context Protocol (MCP) server. It is used to provide context for LLM that supports it. 
+Pass the --stdio flag to use standard input/output for communication. By default, it will bind locally to localhost:8070 (modify it with the --host, --port and --domain flags).
+You can also expose the tunnel to the internet by using the --tunnel flag (useful when the LLM is on a remote server).
+
+
+```bash
+anyquery mcp [flags]
+```
+
+### Options
+
+```bash
+  -c, --config string       Path to the configuration database
+  -d, --database string     Database to connect to (a path or :memory:)
+      --domain string       Domain to use for the HTTP tunnel (empty to use the host)
+      --extension strings   Load one or more extensions by specifying their path. Separate multiple extensions with a comma.
+  -h, --help                help for mcp
+      --host string         Host to bind to (default "127.0.0.1")
+      --in-memory           Use an in-memory database
+      --log-file string     Log file
+      --log-format string   Log format (text, json) (default "text")
+      --log-level string    Log level (trace, debug, info, warn, error, off) (default "info")
+      --port int            Port to bind to (default 8070)
+      --read-only           Open the SQLite database in read-only mode
+      --readonly            Open the SQLite database in read-only mode
+      --stdio               Use standard input/output for communication
+      --tunnel              Use an HTTP tunnel, and expose the server to the internet (when used, --host, --domain and --port are ignored)
+```
+
+### SEE ALSO
+
+* [anyquery](../anyquery)	 - A tool to query any data source

website/src/content/docs/docs/reference/Commands/anyquery_gpt.md

@@ -10,11 +10,12 @@ Create a new profile
 Create a new profile.
 
 If only two arguments are provided, we consider that the registry is the default one.
+If only one argument is provided, the CLI will prompt you for the profile name, and will fallback to the default registry.
 If no argument is provided, the command will prompt you the registry, the plugin and the profile to create.
 Note: This command requires the tty to be interactive.
 
 ```bash
-anyquery profiles new [registry] [plugin] [profile] [flags]
+anyquery profiles new (registry plugin profile) | (plugin profile) | (plugin) [flags]
 ```
 
 ### Examples

website/src/content/docs/docs/reference/Commands/anyquery_mcp.md

@@ -4,12 +4,13 @@ import { $ } from "bun";
 import { readdir } from "node:fs/promises";
 
 // Export the commands
-try {
+/* try {
     await $`anyquery tool generate-doc Commands`;
 } catch (error) {
     // In case it's a build machine, anyquery might not be installed
-    await $`../../../../../../main.out tool generate-doc Commands`;
-}
+    
+} */
+await $`../../../../../../main.out tool generate-doc Commands`;
 
 // For each file, take the lop level h2 header, and add the frontmatter
 // with the title equaling to the header text

website/src/content/docs/docs/reference/Commands/anyquery_profiles_new.md

@@ -22,7 +22,7 @@ Before you begin, ensure that you have the following:
 
 Anyquery supports function calling and the Model Context Protocol (MCP) to connect to your chat interface. Refer to each [connection guide](/integrations#llm) in the documentation for more information.
 
-<LinkCard title="Connection guides" href="/docs/integrations#llm" description="Learn how to connect your chat interface to Anyquery." />
+<LinkCard title="Connection guides" href="/integrations#llm" description="Learn how to connect your chat interface to Anyquery." />
 
 ## Usage
 

website/src/content/docs/docs/reference/export-commands.mjs

@@ -7,7 +7,8 @@ import Main from "../layouts/main.astro";
     footnote="Trademarks are the property of their respective owners. Anyquery is not affiliated with any of the companies or projects mentioned. Some LLM answers may be truncated for brevity."
 >
     <div class="max-w-2xl mx-auto text-center mb-12">
-        <h1 class="text-3xl tracking-[-1.2px] font-bold mt-8">
+        <p class="text-white/70 text-xs mt-8">✨ Anyquery AI ✨</p>
+        <h1 class="text-3xl tracking-[-1.2px] font-bold">
             Connect LLM to your personal data
         </h1>
         <p class="text-white/70 mt-1 text-sm">