Skip to content
The Agents Index
Sign inSubmit an agent

Glean

Enterprise AI platform: permission-aware search, assistant and no-code agents over connected work apps.

Glean is an enterprise AI platform that indexes company apps into a permission-aware knowledge graph and runs search, an assistant and no-code agents on top of it. It exposes that context to outside AI hosts through REST APIs and remote MCP servers.

Best for
Large and mid-market enterprises whose biggest AI-agent blocker is fragmented, permission-sensitive knowledge scattered across 100+ internal tools, and who want employees building department-specific agents on that indexed knowledge without writing code.
Not for
Small teams or budget-conscious buyers who want transparent self-serve pricing (Glean is entirely sales-gated with a six-figure typical deal size), or anyone who wants a single-purpose customer-facing support agent (Sierra, Decagon, Intercom Fin) or a pure workflow-automation canvas (n8n, Dify) rather than a company-wide knowledge-and-agent layer.

What we checked

Pages last read
Facts checked
35 factsFrom 48 sources, read 19 October 2026
AI answers
Named in 3 of 150Our October check

How we checkEditorial policy

Visit Glean CompareAlternatives

Glean in our research

What we checked on Glean

YesPartlyNo

What it doesYes

Writes code

Code Writer is a coding agent that clones connected GitHub repositories, edits them and opens pull requests under the user's own Git identity; it needs a connected code host.

Clones repos, makes changes, runs checks, and opens PRs as you, under your Git identity.

docs.glean.com/security/code-writer-sec…Checked

What it doesPartly

Debugs and fixes

Glean traces a failure through code, PRs, tickets and Slack and Code Writer can propose the fix as a draft pull request, but the vendor frames it as small, targeted changes rather than general repair.

Propose and apply small, targeted code changes by opening draft pull requests in GitHub. Move from bug to fix without leaving Glean.

docs.glean.com/user-guide/assistant/cod…Checked

What it doesPartly

Writes tests

Test changes appear in the vendor's own agent example output; no dedicated test-generation feature is documented.

Implements recurring billing API and background processing according to the documented design, including schema, endpoints, worker, and tests

glean.com/agent-library/spec-to-impleme…

What it doesPartly

Reviews code

Glean supplies enterprise context to an IDE assistant for pull-request review rather than issuing its own review verdicts on a diff, and the vendor describes this as a use case for MCP hosts.

Use Glean context in IDEs (for example, Cursor, VS Code) for debugging, PR reviews, and code navigation.

docs.glean.com/administration/platform/…Checked

What it doesPartly

Whole-codebase work

Code Writer attaches to the right repositories per task, but the vendor scopes its edits to small, targeted changes, so a repository-wide migration is not a claimed capability.

Automatically identifies and attaches to the right repositories for each task, using your GitHub integration and configuration.

docs.glean.com/security/code-writer-sec…Checked

What it doesNo

Ships it

The coding agent stops at a pull request and defers to existing branch protections, CODEOWNERS and approval flows; no vendor page claims Glean deploys or releases the change.

Fully respects your existing branch protections, CODEOWNERS, and approval flows.

docs.glean.com/security/code-writer-sec…Checked

What it doesNo

Done-for-you service

Buyers license software per user; Glean operates the infrastructure and application in either deployment model but the vendor's pages describe no consulting or build-for-you engagement.

Glean is responsible for infrastructure and application management, including upgrades, patching, and monitoring.

docs.glean.com/get-started/prepare/abou…Checked

How much it does unattendedYes

Runs autonomously

Autonomous agents plan their own approach from plain-language instructions and time-based triggers run them in the background, within tool-call and run-time guardrails.

Autonomous agents interpret your plain-language instructions to determine the best approach to solve complex enterprise use cases

glean.com/ai-agents/agent-builderChecked

How much it does unattendedYes

Multi-agent

An orchestration layer routes requests to specialist agents and chains them through multi-step work, passing context between the agents it activates.

Orchestration connects specialized agents to complete multi-step tasks, activating each one at the right time and sharing context seamlessly along…

glean.com/ai-agents/agent-orchestrationChecked

How much it does unattendedPartly

Agent permissions

Write tools ask for confirmation unless an admin marks them eligible and the agent builder also enables it for that step, so unattended write access is a deliberate two-level opt-in.

This admin setting does not make the tool run automatically or change an existing Agent. An Agent builder must also configure the relevant step to…

docs.glean.com/administration/tools/man…Checked

Which models it runs onYes

Claude

Anthropic models are part of the Model Hub's curated set and Claude Haiku, Sonnet and Opus variants appear with their own token rates in the published rate card.

The Model Hub offers a curated set of leading models from providers like OpenAI, Google, Anthropic, and Amazon.

docs.glean.com/administration/llmsChecked

Which models it runs onYes

GPT

OpenAI is a listed Model Hub provider and the rate card prices GPT-4o, GPT-4.1 and GPT-5 families per million tokens, so GPT is selectable for Assistant and agents.

On the page: Supported LLMs, opening paragraph

docs.glean.com/administration/llmsChecked

Which models it runs onYes

Open models

Open-weight models are served through Glean's providers under zero-retention terms, and admins enable them by creator region rather than one model at a time.

Open models are models with openly published weights, such as GLM 5.2 and Nemotron 3 Ultra.

docs.glean.com/administration/open-mode…Checked

Which models it runs onYes

Model choice

Admins toggle which of the 40-plus models are available and can leave selection to auto routing, and the same hub supports bring-your-own provider keys.

Glean’s Model Hub gives teams 40+ open and proprietary LLMs, with auto routing and controls for access, safety, and spend.

glean.com/platform/model-hubChecked

Where you use itYes

In your editor

Glean's remote MCP server plugs into IDE hosts including Cursor, VS Code and Windsurf, and plug-ins exist for Cursor, Claude Code and Codex.

Developers and power users can access Glean's search, chat, and agents directly in preferred tools such as Cursor, VS Code, Claude Desktop, and…

docs.glean.com/administration/platform/…Checked

Where you use itPartly

On the command line

There is no Glean CLI of its own; terminal use comes from plug-ins and MCP connections inside CLI agents such as Claude Code, Codex and Gemini CLI.

Glean also provides plug-ins for Cursor, Claude Code, and Codex. The plug-ins discover skills and tools dynamically and run them through Glean's…

docs.glean.com/administration/platform/…Checked

Where you use itPartly

In your pipeline

Agents act on pull requests through the GitHub API, writing comments, descriptions and patches, but Glean runs no CI job of its own and respects existing branch protections.

Using the GitHub API, the agent adds comments directly to the PR.

glean.com/agent-library/pull-request-re…Checked

Where you use itYes

In a browser

The primary surface is a hosted web app, alongside a desktop app, browser extension and Slack, with MCP carrying Glean into other AI tools.

Open Glean in the browser, desktop app, browser extension, Slack, and other places your day already happens.

docs.glean.com/user-guide/about/what-is…Checked

Whose machine it runs onPartly

Self-hosted

Customer Hosted puts the single-tenant deployment inside the buyer's own AWS or GCP account, but Glean still deploys and manages it as a service rather than handing over software.

Glean deploys its tenant as a managed service in isolation within your own cloud environment (GCP, AWS). Previously known as Cloud-Prem.

docs.glean.com/get-started/prepare/abou…Checked

Whose machine it runs onNo

Open source

Closed, vendor-operated system: even in the customer's own cloud, altering or patching Glean services is unsupported.

In the Customer Hosted model, Glean doesn't support manually deploying or patching the Glean services, or altering any part of the Glean architecture.

docs.glean.com/get-started/prepare/abou…Checked

What it costs to runYes

How it meters

Two published models: per-user monthly seats, plus metered model consumption billed as FlexCredits or Model Hub usage priced per million tokens on top of the seat fee.

All prices are per million tokens and sourced from each provider's API pricing

docs.glean.com/glean-core-suite-pricingChecked

What it costs to runNo

Free tier

Licensed per user per month with an upfront usage commit; every call to action on the site is a demo request, and no self-serve or free tier is published.

Licensed on a per-user, per-month basis, Glean Core Suite includes:

docs.glean.com/glean-core-suite-pricingChecked

What it costs to runYes

API access

Documented REST APIs and SDKs cover search, chat, agents and indexing, with Python, Go, Java and TypeScript clients; LLM-backed calls are billed at Model Hub usage rates.

Glean provides APIs and tools for search, chat, agents, indexing, web embedding, remote MCP, and custom actions.

glean.com/platform/apiChecked

What it costs to runYes

MCP server

A remote MCP server exposes Glean search, chat, document read and agents as tools to external hosts, with 22 documented hosts and OAuth 2.1 authentication.

Bring Glean's permission-aware enterprise context into your preferred MCP hosts.

docs.glean.com/administration/platform/…Checked

What it costs to runYes

Bring your own key

Customers either use Glean's Universal Model Key or bring their own provider keys, and Customer Hosted additionally runs the whole deployment in the buyer's cloud account.

You can configure models using either Glean Universal Model Key or Customer Key (BYOK).

docs.glean.com/administration/configure…Checked

Buying it for a teamYes

A company can buy it

This is sold to organisations only: licensing is per user per month across two named enterprise plans, and single sign-on is mandatory before anyone can log in.

On the page: Glean Core Suite, paragraph above the feature table

docs.glean.com/glean-core-suite-pricingChecked

Buying it for a teamYes

Seat model

Seats are per user per month and the Enterprise Flex plan expects them for every employee, which is a practical floor; no seat figure or explicit minimum count is published.

Licensed on a per-user, per-month basis, Enterprise Flex Seats are designed to be deployed to every employee in your organization

docs.glean.com/glean-enterprise-flex-pr…Checked

Buying it for a teamYes

Pooled budget

Enterprise Flex bundles a shared credit allowance with the seat licences and bills only above it; admins can also cap spend per organisation, user, team or agent monthly.

Glean includes a pool of credits as part of the per-user licenses and you only pay when you go above the initial allocation.

docs.glean.com/glean-enterprise-flex-pr…Checked

Buying it for a teamYes

Admin controls

Admins set agent sharing scope, per-tool write eligibility, restricted topics and which models each department may use, and agents run with the signed-in user's permissions.

Admins can define how broadly agents can be shared, while builders choose whether to share their agents with individuals, teams, or the whole company…

glean.com/ai-agents/agent-governanceChecked

Buying it for a teamYes

Audit log

Admin audit logs record configuration and administrative actions with filtering and CSV export, but they deliberately exclude end-user activity, which lives in Insights instead.

Glean's Admin Audit Logs provide comprehensive visibility into administrative actions performed within your Glean instance.

docs.glean.com/administration/managemen…Checked

Buying it for a teamYes

Single sign-on

Both SAML and OIDC are supported with Okta, Entra ID, Google and generic SAML providers, and single sign-on is required rather than optional on this product.

Glean SSO utilizes standard authentication protocols such as OpenID Connect (OIDC) and Security Assertion Markup Language (SAML)

docs.glean.com/administration/identity/…Checked

What happens to your codePartly

Opt out of training

On Glean's own model key the vendor states customer data is not used for model training and open models run under zero-retention terms; with bring-your-own keys the provider contract governs instead.

Get ZDR-compliant access to leading open and frontier models across providers. Your data isn’t used for model training, with consistent security…

glean.com/platform/model-hubChecked

What happens to your codeYes

Data residency

Preferred hosting regions are named in the US (Iowa), APAC (Taiwan) and EMEA (Netherlands), with other GCP and AWS regions supported for data-sovereignty requirements; tenants default to us-central1.

Glean recommends leveraging a preferred region for hosting your Glean tenant unless a specific country is required for data sovereignty purposes.

docs.glean.com/get-started/prepare/self…Checked

What happens to your codePartly

Getting out

Usage data, audit logs and org charts export as CSV and GCE event logs support custom reporting, but no vendor page states whether unused FlexCredits survive the contract period.

You can review overall usage trends, break down consumption by feature and surface, and export data for your own reports.

docs.glean.com/administration/managemen…Checked

What happens to your codeYes

Certifications

The security page displays SOC 2 Type II, ISO 27001, ISO 42001, HIPAA, GDPR and TX-RAMP Level 2, and healthcare pages repeat the HIPAA claim in prose.

Keep data safe and audit-ready with HIPAA compliance and permission-aware access

glean.comChecked

Glean price and terms

What needs a higher plan

Enterprise Flex
Pooled budget

For AI agents

MCP server
Yes
API
Yes

Developer docs

Glean's pricing page

Pricing page, captured

glean.com/pricing, checked

AdvertisementReach the teams comparing AI agent platforms here.One banner on The Agents Index's category pages and agent pages.Advertise here

Glean alternatives

All 9 alternatives

Head to headGlean vs Moveworks

Questions about Glean

How much does Glean cost?
Glean gives its price on request. From glean.com/pricing, checked 28 October 2026.
Does Glean have an API or an MCP server?
It has an API. It has an MCP server.
Is Glean open source, and can you host it yourself?
It is not open source. You can host part of it yourself.
What have we checked on Glean?
35 facts, under these headings: What it does, how much it does unattended, which models it runs on, where you use it, whose machine it runs on, what it costs to run, buying it for a team, what happens to your code.
How do we check Glean?
We hold 35 facts on Glean, quoted from 48 pages. Facts last checked 19 October 2026. Price last checked 28 October 2026.

How we check every listing

Sources and changes

Last changed

Sources48

Written about it elsewhere4

Is this yours?

AI assistants like ChatGPT and Claude opened The Agents Index pages 2,746 times in September. Claiming is free. You get this page's reader numbers and the verified owner mark. You confirm it with an email at glean.com.

Claim this listingGet Featured

Report an error if a fact is wrong.