Enables AI agents to perform search, schema inspection, collection operations, and cluster administration on Typesense databases.
Also known as akshitkrnagpal/typesensekit, @typesensekit/mcp
- Stars
- 3
- Tools
- —
- Downloads, 30 days
- —
- Latest version
- 1.4.3
- Last commit
- Nov 10, 2026
- License
- MIT
What is TypesenseKit MCP?
TypesenseKit enables AI agents to perform search, schema inspection, collection operations, and cluster administration on Typesense databases.
Summary generated from the project's README
About TypesenseKit MCP server
TypesenseKit provides AI agents with a structured interface to interact with Typesense databases. It allows agents to perform search queries, inspect schemas, manage collections, and handle cluster administration tasks. The server runs in a read-only mode by default to ensure safety, but users can enable write, delete, and admin capabilities by setting the TYPESENSEKIT_READ_ONLY environment variable to false.
Developers can deploy this as a local tool or run it as a stateless Streamable HTTP server using the typesensekit-mcp-http package. Connection requires a TYPESENSE_URL and a TYPESENSE_API_KEY. To manage limited context windows in AI clients, the server includes a compact mode for the collections.list tool. This mode returns essential metadata like document counts, aliases, and field types while omitting heavy schema details. It also offers a compact search option to reduce response size when highlights and match scores are not required.
What you can do with it
- 1Optimized Context ManagementA developer can instruct an agent to list collections using a compact format to save tokens.
- 2Efficient Document SearchingAn agent can perform searches while requesting only specific fields to minimize the amount of data returned.
- 3Full Database AdministrationAn agent can be granted permission to modify or delete data within the Typesense cluster.
Written from the project's README. Every use case is checked against a quote from it.
How to install TypesenseKit MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall TypesenseKit MCP in Claude Code
$ claude mcp add --env 'TYPESENSE_API_KEY=<YOUR_TYPESENSE_API_KEY>' --env 'TYPESENSE_URL=<TYPESENSE_URL>' typesensekit -- npx -y @typesensekit/mcpInstall TypesenseKit MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"typesensekit": {
"command": "npx",
"args": [
"-y",
"@typesensekit/mcp"
],
"env": {
"TYPESENSE_API_KEY": "<YOUR_TYPESENSE_API_KEY>",
"TYPESENSE_URL": "<TYPESENSE_URL>"
}
}
}
}Install TypesenseKit MCP in VS Code
.vscode/mcp.json
{
"servers": {
"typesensekit": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@typesensekit/mcp"
],
"env": {
"TYPESENSE_API_KEY": "<YOUR_TYPESENSE_API_KEY>",
"TYPESENSE_URL": "<TYPESENSE_URL>"
}
}
}
}Install TypesenseKit MCP in VS Code CLI
$ code --add-mcp '{"name":"typesensekit","type":"stdio","command":"npx","args":["-y","@typesensekit/mcp"],"env":{"TYPESENSE_API_KEY":"<YOUR_TYPESENSE_API_KEY>","TYPESENSE_URL":"<TYPESENSE_URL>"}}'Install TypesenseKit MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"typesensekit": {
"command": "npx",
"args": [
"-y",
"@typesensekit/mcp"
],
"env": {
"TYPESENSE_API_KEY": "<YOUR_TYPESENSE_API_KEY>",
"TYPESENSE_URL": "<TYPESENSE_URL>"
}
}
}
}Install TypesenseKit MCP in Codex
$ codex mcp add typesensekit --env 'TYPESENSE_API_KEY=<YOUR_TYPESENSE_API_KEY>' --env 'TYPESENSE_URL=<TYPESENSE_URL>' -- npx -y @typesensekit/mcpInstall TypesenseKit MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.typesensekit] command = "npx" args = ["-y", "@typesensekit/mcp"] [mcp_servers.typesensekit.env] TYPESENSE_API_KEY = "<YOUR_TYPESENSE_API_KEY>" TYPESENSE_URL = "<TYPESENSE_URL>"
Install TypesenseKit MCP in Gemini CLI
$ gemini mcp add -s user -e 'TYPESENSE_API_KEY=<YOUR_TYPESENSE_API_KEY>' -e 'TYPESENSE_URL=<TYPESENSE_URL>' typesensekit npx -- -y @typesensekit/mcpInstall TypesenseKit MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"typesensekit": {
"command": "npx",
"args": [
"-y",
"@typesensekit/mcp"
],
"env": {
"TYPESENSE_API_KEY": "<YOUR_TYPESENSE_API_KEY>",
"TYPESENSE_URL": "<TYPESENSE_URL>"
}
}
}
}Configuration
TYPESENSE_API_KEYRequiredSecretTypesense API key used to authenticate server operations.
TYPESENSE_URLRequiredURL of the Typesense server, including protocol and port when required.
TYPESENSEKIT_READ_ONLYExpose only read operations unless explicitly disabled.
TYPESENSE_CONNECTION_TIMEOUT_SECONDSOptional Typesense connection timeout in seconds.
| Name | Description | |
|---|---|---|
| TYPESENSE_API_KEY | Typesense API key used to authenticate server operations. | RequiredSecret |
| TYPESENSE_URL | URL of the Typesense server, including protocol and port when required. | Required |
| TYPESENSEKIT_READ_ONLY | Expose only read operations unless explicitly disabled. | |
| TYPESENSE_CONNECTION_TIMEOUT_SECONDS | Optional Typesense connection timeout in seconds. |
Automated checks
Ran Nov 13, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablePass
- Open source license (OSI)Pass
- Commit or release in the last yearPass
- Package name not close to a popular packagePass
- Remote endpoint uses HTTPSNot checked
- Destructive tools are labelledNot checked
- No hidden instructions to the modelPass
- No critical advisory in the package or its dependenciesPass
README
View on GitHub@typesensekit/mcp
An open-source Typesense MCP server that gives AI agents structured tools for search, schema inspection, collection operations, and cluster administration. It runs in read-only mode by default.
Typesense MCP overview · Client setup · Security guide · GitHub
TYPESENSE_URL=http://localhost:8108 TYPESENSE_API_KEY=xyz pnpm dlx @typesensekit/mcp
Set TYPESENSEKIT_READ_ONLY=false to expose write/delete/admin tools.
Run the stateless Streamable HTTP server:
TYPESENSE_URL=http://localhost:8108 TYPESENSE_API_KEY=xyz pnpm dlx --package @typesensekit/mcp typesensekit-mcp-http
See the API coverage inventory for supported Typesense operations.
For smaller agent contexts, use collections.list with {"compact":true}.
It returns {collections}: per collection the document count, aliases,
default sorting field, linked synonym/curation sets, text_fields (indexed
string fields to use in query_by), and fields as
name: type [flags] strings that list only non-default flags (for example
price: float [facet, sort] or notes: string [index:false, optional]).
Aliases whose target collection is missing are returned under aliases.
This mode reads both /collections and /aliases; the API key needs access
to both. Use collections.retrieve for full field definitions. Omitting
compact keeps the full collection schema array.
Tool listings are compact: they omit $schema, additionalProperties: false
and descriptions that repeat the server instructions, and name rarely used
options (vector settings on collections.fields.*) in the description
instead of typing them. Calls are still validated against the full input
schemas, so every documented input is accepted. Errors about a field's
schema flags (facet, sort, index, missing query_by) list the collection's
fields and text fields so the request can be fixed without another lookup.
For document searches, prefer search with compact:true and
params.include_fields when highlights and match scores are unnecessary.
documents.search remains an equivalent alias. Search output defaults are
unchanged; JSON text responses use no indentation and preserve the same data.
TypesenseKit MCP questions
How do I add TypesenseKit MCP to Claude Code?
Run claude mcp add --env 'TYPESENSE_API_KEY=<YOUR_TYPESENSE_API_KEY>' --env 'TYPESENSE_URL=<TYPESENSE_URL>' typesensekit -- npx -y @typesensekit/mcp in your terminal. Add --scope user to make it available in every project.
How do I add TypesenseKit MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
Is the server read-only by default?
Yes, the server runs in read-only mode by default to ensure safety.
How can I enable write and delete operations?
You can expose write, delete, and admin tools by setting the environment variable TYPESENSEKIT_READ_ONLY to false.
How can I reduce the context size for smaller agent contexts?
Use the collections.list tool with the compact parameter set to true to receive a condensed version of the collection data.