Skip to main content

Glean Developer Platform

Glean Platform APIsGenerally available

  • Search (generally available)
  • Chat (generally available)
  • Agents (generally available)
  • Skills (experimental)
  • Triggers (experimental)
Explore Platform APIs
Chat API

Chat grounded in company knowledge

Ask across every connected app and get permission-aware answers, with citations, in a single API call.

Start buildingAPI reference
7 recipes in the Cookbook
chat.pyChat API
from glean.api_client import Glean

with Glean(
    api_token=os.environ["GLEAN_API_TOKEN"],
    server_url=os.environ["GLEAN_SERVER_URL"],
) as client:
    response = client.client.chat.create(
        messages=[{"fragments":
            [{"text": "Summarize the Q3 roadmap"}]}]
    )
    answer = "".join(
        f.text or ""
        for m in response.messages or []
        for f in m.fragments or []
    )
search.pySearch API
from glean.api_client import Glean

with Glean(
    api_token=os.environ["GLEAN_API_TOKEN"],
    server_url=os.environ["GLEAN_SERVER_URL"],
) as glean:
    results = glean.client.search.query(
        query="quarterly reports",
        page_size=10,
    )
    titles = [r.title for r in results.results]
agent.pyAgents
from glean.agent_toolkit.tools import (
    search,
    employee_search,
)

# Use with LangChain
search_tool = search.as_langchain_tool()
people_tool = employee_search.as_langchain_tool()

# Use with CrewAI
crew_tool = search.as_crewai_tool()
embed.tsWeb SDK
import {
  renderSearchBox,
  renderSearchResults,
  renderChat,
} from "@gleanwork/web-sdk";

renderSearchBox(searchEl, {
  backend: "https://acme-be.glean.com",
  onSearch: (query) =>
    renderSearchResults(resultsEl, { query }),
});

renderChat(chatEl, {
  backend: "https://acme-be.glean.com",
});
connector.pyIndexing SDK
from glean.indexing.connectors import (
    BaseDatasourceConnector,
)

class CatalogConnector(BaseDatasourceConnector):
    def get_data(self):
        return load_catalog_pages()

connector = CatalogConnector(name="catalog")
connector.index_data(mode=IndexingMode.FULL)

Choose your path

Four ways to get started.

Build with Platform APIStart with modern search, agents, skills, and chat from your code.Embed with the Web SDKPermission-aware search and chat inside the apps your team already uses.
Connect your dataBring any source into Glean with the Indexing API.Bring Glean to your IDEClaude Code, Cursor, Codex, and any MCP host.
Quickstart

One call to your knowledge graph

  1. 1Prefer OAuth for per-user Client and Platform work in Authentication. Use a Glean-issued token for Indexing, global ActAs, or when no OAuth path exists.
  2. 2Install a client library for your language.
  3. 3Run your first query — results come back ranked and permission-aware.
main.py
from glean.api_client import Glean

with Glean(
    api_token=os.environ["GLEAN_API_TOKEN"],
    server_url=os.environ["GLEAN_SERVER_URL"],
) as glean:
    results = glean.client.search.query(
        query="quarterly planning",
        page_size=10,
    )
    for r in results.results:
        print(r.title, r.url)

API client libraries

Python
TypeScript
Java
Go

Bring Glean into your IDE

Claude CodeInstall the Glean plugins and search company knowledge from your terminal. CursorConnect the Glean MCP server to Cursor for context-aware coding.CodexCodexInstall the Glean plugin for Codex — enterprise knowledge in your terminal.

Plus GitHub Copilot, Goose, Windsurf, and any MCP host.

Agents

Framework-agnostic by design

The agent toolkit exposes Glean retrieval as tools for whatever you build with — the knowledge graph comes along regardless of framework.

LangChainOpenAI Agents SDKGoogle ADKMCP
Explore the agent toolkit
tools.pyAgent toolkit
from glean.agent_toolkit.tools import (
    search,
    employee_search,
)

# LangChain
lc_tool = search.as_langchain_tool()

# CrewAI
crew_tool = search.as_crewai_tool()