Skip to main content
hc
Submit
mcpSource-backedReview first Safety ✓ Privacy ✓

Algolia MCP Server for Claude

Connect Claude to Algolia-managed MCP servers for curated public search and recommendation access or internal, user-scoped search and analytics workflows.

by Algolia·added 2026-06-05·
Claude CodeClaude Desktop
HarnessClaude CodeClaude Desktop
Review first — review before installing

Open the source and read safety notes before installing.

Safety notes

  • Algolia documents two Algolia-managed MCP offerings. Public MCP is for curated search and recommendation access to selected indices; Productivity MCP is for internal, user-scoped workflows that inherit the signed-in user's Algolia permissions.
  • Algolia describes Public MCP as a beta feature under its Beta Services terms. Treat setup behavior, limits, dashboard controls, and client support as changeable until Algolia marks the feature generally available.
  • Public MCP uses a generated server URL for a specific Algolia application and selected indices. Do not publish that URL in public repositories, issues, logs, screenshots, or client configs unless the exposed records are intended for that audience.
  • Public MCP client configuration does not require client-side authentication, so the selected Search API key restrictions, selected indices, Recommend models, and index descriptions are the primary controls.
  • Productivity MCP is documented as read-only for analysis and exploration, but it can expose anything the signed-in user can access in Algolia. Use a least-privilege account for sensitive projects.
  • Algolia says MCP usage counts toward the existing Algolia plan. Monitor API usage and avoid broad automated loops, high-traffic prompts, or repeated large searches through an MCP client.

Privacy notes

  • Algolia MCP can return search hits, object attributes, facet values, recommendations, index names, analytics metrics, top searches, no-result queries, countries, click positions, user counts, and other search performance data into the model conversation.
  • Public MCP exposes records according to the configured indices and Search API key permissions. Remove private object attributes or use secured, filtered, or separate indices before connecting customer-facing agents.
  • Productivity MCP uses the signed-in user's Algolia identity and permissions. Treat prompts, outputs, screenshots, and MCP client logs as internal data whenever they include Algolia analytics, index names, or object records.
  • Do not paste Admin API keys, write API keys, private secured API keys, customer identifiers, unreleased product catalog fields, pricing rules, personal data, or raw exports into prompts or PR comments.
  • Algolia recommends prompt guidance that preserves result fidelity. Avoid asking the assistant to filter, reorder, or summarize returned hits when an exact search result order matters.

Prerequisites

  • Algolia account with access to the applications and indices you want Claude to use.
  • MCP-capable client with remote HTTP transport support, such as Claude, Claude Code, ChatGPT, Cursor, VS Code, Gemini CLI, Codex, OpenAI Playground, or another compatible client.
  • For Algolia Productivity MCP, access to enable Productivity MCP in the Algolia dashboard and sign in through the MCP client when prompted.
  • For Algolia Public MCP, an Algolia application with at least one index and the dashboard permission to create or copy a Public MCP server URL.
  • Decision on whether the workflow should use Public MCP for curated application-scoped access or Productivity MCP for internal user-scoped analysis.
  • Review of Algolia plan usage, Search API key permissions, index descriptions, and prompt/result handling before exposing production indices to AI clients.

Schema details

Install type
cli
Troubleshooting
No
Collection metadata
Estimated setup
10 minutes
Difficulty
intermediate
Tool listing metadata
Full copyable content
{
  "mcpServers": {
    "algolia": {
      "transport": "http",
      "url": "https://mcp.algolia.com/mcp"
    }
  }
}

About this resource

Content

Algolia MCP Server is Algolia's managed Model Context Protocol support for connecting AI assistants to Algolia search and analytics workflows. Algolia documents two managed offerings:

  • Algolia Public MCP exposes selected indices and Recommend models through an application-scoped MCP URL for AI-facing search and recommendation use cases.
  • Algolia Productivity MCP exposes internal, user-scoped search and analytics tools through https://mcp.algolia.com/mcp after the user signs in.

Use Public MCP when the assistant should search a curated set of records, such as product discovery, documentation search, support answers, or customer-facing AI experiences. Use Productivity MCP when a team member wants Claude to inspect top searches, no-result queries, click positions, countries, users, or index data that the signed-in Algolia account can already access.

Features

  • Algolia-managed MCP services, not a self-hosted integration requirement.
  • Public MCP for selected indices and Recommend models.
  • Productivity MCP for internal, user-scoped search and analytics exploration.
  • Claude custom connector setup for both Public MCP and Productivity MCP.
  • Claude Code setup for Productivity MCP with remote HTTP transport.
  • Codex, Cursor, VS Code, Gemini CLI, ChatGPT, and OpenAI Playground setup paths documented by Algolia.
  • Public MCP URL pattern: https://{APP_ID}.algolia.net/mcp/1/{UNIQ_ID}/mcp.
  • Productivity MCP URL: https://mcp.algolia.com/mcp.
  • Public MCP tools for search, facet values, recommendations, and ChatGPT search/fetch compatibility.
  • Productivity MCP tools for listing/searching indices, recommendations, and analytics such as top searches, no-result queries, click positions, filters, countries, users, and top hits.
  • Existing Algolia plan usage applies to MCP calls.

Installation

Productivity MCP for Claude Code

Enable Algolia Productivity MCP in the Algolia dashboard, then register the remote MCP server:

claude mcp add --transport http algolia https://mcp.algolia.com/mcp

Start a Claude Code session, run /mcp, and sign in when prompted to authorize Algolia.

Productivity MCP Client Config

Use this remote HTTP configuration for MCP clients that accept JSON configuration:

{
  "mcpServers": {
    "algolia": {
      "transport": "http",
      "url": "https://mcp.algolia.com/mcp"
    }
  }
}

The client should prompt for OAuth-style sign-in. Algolia's docs say no OAuth client ID or client secret is needed when the client asks for those fields.

Public MCP Client Config

Create or copy a Public MCP server URL from the Algolia dashboard, then add it to your MCP client:

{
  "mcpServers": {
    "Algolia": {
      "url": "https://{APP_ID}.algolia.net/mcp/1/{UNIQ_ID}/mcp"
    }
  }
}

For Claude, add the copied URL as a custom connector remote MCP server URL and leave advanced settings empty.

Use Cases

  • Let a customer-facing assistant search only the curated product, help center, or documentation indices selected for Public MCP.
  • Ask Claude to return Algolia hits exactly as provided for a search or support workflow.
  • Explore top searches and no-result queries before planning synonyms, rules, content fixes, or index changes.
  • Review click positions, no-click rate, top countries, top filters, and top hits while diagnosing search relevance.
  • Validate which indices and attributes are visible before adding a new AI search experience.
  • Compare recommendation results from selected Recommend models without building a separate integration.
  • Use Productivity MCP for one-off internal analysis while keeping production updates in the Algolia dashboard or APIs.

Safety Checklist

  • Choose Public MCP for a deliberately curated external-facing surface.
  • Choose Productivity MCP only for internal users whose Algolia permissions are appropriate for the data being queried.
  • For Public MCP, select only the indices and Recommend models that the agent should use.
  • Add descriptions that help the LLM choose the right index or tool.
  • Confirm Search API key restrictions before exposing a Public MCP URL.
  • Keep generated Public MCP URLs out of public code, logs, tickets, and screenshots.
  • Monitor Algolia usage because MCP calls count toward the existing plan.
  • Use prompt instructions that preserve Algolia result order and content when exact search results matter.

Source Links

#algolia#search#recommendations#analytics#mcp

Source citations

Signals

Loading live community signals…

More like this, weekly

A short, calm digest of reviewed Claude resources. Unsubscribe any time.