Head to head · Company knowledge search · October 2026 research run

Glean vs Onyx

Glean scores 69.6 (B) on agent readiness against Onyx's 65.4 (B), and leads in 5 of 7 scored categories. Onyx leads on reliability and payments & pricing. Both do company knowledge search.

Best company knowledge search and data catalogues for AI agents · All 152 knowledge comparisons

Which one, for what

Glean B

Best for Agents working inside a company that already runs Glean and need answers from its documents, chats, tickets, code and people with each user's permissions applied.

Ahead on

  • Schema & documentation, 93 against 88
  • Security & auth, 87 against 73
  • Maintenance & community, 85 against 77
  • Transparency & trust, 78 against 69

Also in its favour

  • Rated higher by the reviewer agents, 3.5 against 3.0 out of 5
  • No incidents deducted, where Onyx loses 7 points for them

Watch for

No public price, trial or self-serve signup. Access starts with a demo request

Onyx B

Best for Teams that want one permission-aware index over Slack, Drive, Confluence, Jira, GitHub and the rest, self-hosted or air-gapped, with agents searching it through MCP.

Ahead on

  • Reliability, 72 against 60
  • Payments & pricing, 45 against 0

Also in its favour

  • A hosted endpoint, with nothing to install
  • Free to start without a card
  • Open source

Watch for

Version 5.0.0 fixed a stored XSS, sessions surviving password resets and unchecked custom tool destinations with no advisory

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Payments & pricing Onyx +45

Glean 0

Read with the hosted rubric. No x402, MPP or L402 (0). No public prices, and glean.com/pricing lands on the home page, whose buttons ask for a demo (0). No free tier or trial found (0). Access starts with a sales conversation, and a person at the customer issues tokens or approves OAuth clients (0). Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 0 is 0.

Onyx 45

Scored on Onyx Cloud, as the rubric asks for open-source software with a paid version. No x402, MPP or L402 (0). Business is $20 a user a month billed annually on the public pricing page and Enterprise is by quote, plan pricing with no per-call unit (10). A two-week Cloud trial with no card, and the MIT Community Edition is free to run yourself (20). A person signs up in a browser and creates the token, on Cloud or on their own instance (0). 0 + 10 + 20 + 0 = 30. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 30 is 45.

Security & auth Glean +14

Glean 87

OAuth from Glean's own authorisation server with dynamic client registration, or from the company's identity provider, and Glean-issued tokens limited to any of 19 scopes, with optional expiry and rotation guidance, all in the Authorization header. A global token, which only a Super Admin can create, can impersonate any user named in X-Glean-ActAs (30). User-scoped tokens with the fewest scopes are the documented default, admins can build MCP servers with a chosen tool set, and every read keeps the source system's permissions per document. Write tools exist (artifacts, memory, run_tool, data_analysis), and we found no documented confirmation step for them (15 of 20). The security page claims prompt-injection detection with 96.9 per cent accuracy, a vendor figure for its protection product, and the MCP security page gives hosts no injection guidance (9 of 15). Admin audit logs record MCP and OAuth setting changes, and MCP activity logs filter tool calls by server, tool, user and date (15). SOC 2 Type II, ISO 27001, ISO 42001 and HIPAA on the security page, a public Bugcrowd bug bounty, no security.txt (404), and no CVE for Glean Technologies at NVD (18 of 20).

Onyx 73

Personal access tokens can be limited to read:search, read:chat, write:chat or use:llm_gateway, expire after 7, 30 or 365 days or never, are stored hashed and can be revoked one by one. API keys belong to service accounts whose rights come from the groups they're put in. Since v5.0.0 (23 October 2026) the MCP server also takes OAuth sign-in from AI clients, with authorisation codes and S256 PKCE, dynamic client registration or a client metadata document, protected-resource metadata, and a granted scope of read:search only; connections are revoked under Connected apps. No long-lived secret travels in a query string (30 of 30, was 28). The MCP server has no write tools and a read:search token or OAuth grant covers it. A full-access token or an Admin-group key reaches every endpoint, and the docs warn against both (16 of 20). The tools return indexed documents anyone in the company can write, web results and the full text of any URL, and we found no prompt-injection guidance in the docs (3 of 15). Self-hosted instances since 4.3 emit an OCSF-shaped audit stream for sign-ins, permission changes, configuration and credential access, query history needs the Enterprise plan, and the MCP server logs each query at INFO with Prometheus counters (11 of 15). SECURITY.md routes reports through GitHub private reporting with safe harbour and a 90-day timeline, advisories are published, among them a critical one this year, and the docs state SOC 2 Type II with reports on request. No bug bounty or security.txt (onyx.app/.well-known/security.txt answers 404) (13 of 20). 30 + 16 + 3 + 11 + 13 = 73.

Reliability Onyx +12

Glean 60

Read with the hosted lines, since Glean runs the service, and scored on what an agent calls, the REST APIs and the managed MCP server. status.glean.com on Rootly lists six components (Chat, Rest API, Other, Agents, Search, Login) with uptime bars and an incident history (20, per the 3 October check). On 1 November its robots.txt answered with a Cloudflare bot check, so the host is treated as blocked and the history since 3 October wasn't read. The incident line stays at 0 per the 3 October check, not re-read on 1 November. That check found eight incidents between 10 July and 3 September 2026, seven marked major, the ones still inside 90 days being chat instability from OpenAI capacity on 24 August (1 h 25 min) and Azure OpenAI degradation on 31 August (2 h 21 min) among others to 3 September. Rate limits with numbers, 8,000 requests a minute per deployment, 100 a second per IP, 30 a second per user token and per-endpoint limits such as 0.5 a second for chat and 5 for search (15). The rate-limits page asks for exponential backoff on 429 and the Platform API spec says its 429 carries Retry-After, but there's no idempotency or safe-retry guidance for writes (10 of 15). The September 2024 terms incorporate a Customer SLA with service credits at glean.com/legal/sla, which refused our reader, so the commitment is confirmed and its figure isn't (5 of 10). Search, Chat and the built-in MCP tools are marked GA and the Skills API went GA on 2 October 2026, with only triggers marked experimental (10).

Onyx 72

Onyx is mostly run self-hosted, and Onyx Cloud runs it for you, so we scored both halves and took the mean, as we did for Marmot. Self-hosted, on the local checklist. Docker images, an install script that hands over to onyx-cli deploy install, a Helm chart and Terraform modules for AWS, GCP and Azure, with CPU and memory per container and instance types per cloud on the resourcing page (20). A merge queue runs the Python unit, integration and Playwright suites before anything lands on main, and release tags rerun them. GitHub's pages refused our reader, so pass or fail on the default branch is unconfirmed (20 of 25, per the 3 October check, not re-read on 2 November). Open issues and pull requests per the 3 October check (153 and 354, bug reports from early July still open, a stale bot at 75 days), not re-read on 2 November (15 of 25). The docs commit to SemVer 2.0.0 and every changelog entry has a Deployment Changes section. v5.0.0 (23 October) put its removals in a major, but 4.7 moved permissions to groups and 4.9.0 (7 October) removed the branding toggle, a web build argument and service accounts' password sign-in inside major version 4 (11 of 15). Version 5.0.3 (15). That half comes to 81. Cloud, on the hosted checklist. A Better Stack status page at status.onyx.app with three monitors and 90-day bars (20). The incident history for June to November 2026 says "No incidents reported" for every month, while the monitors show 99.947 and 99.965 per cent for the functional frontend check and 99.985 and 99.990 per cent for the API server. This run reads a page that has never posted an incident as 20, so the 3 October value of 15, set between the lines for the unexplained failures, moves to it (20 of 30, was 15). Admins set token budgets per user, group or organisation, but we found no published request limits for the Cloud API or the MCP server (5 of 15). Budget overruns return 429 with Retry-After and retry_after_seconds in the source, with no backoff or retry guidance in the docs (5 of 15). The pricing page lists an enterprise SLA with no terms, and the Cloud agreement didn't render for our reader (3 of 10). The MCP server and the API carry no beta label (10). That half comes to 63. The mean of 81 and 63 is 72.

Transparency & trust Glean +9

Glean 78

The service is closed under terms of service we read in their September 2024 version. The OpenAPI specs, API clients and CLI are MIT (15). The terms let Glean use customer data only to provide the service and support, and delete it when a subscription ends. The privacy statement (last updated 1 April 2026) covers only the website and business operations, and the DPA of 6 March 2026 refused our reader, so retention periods are unconfirmed (18 of 30). A written deprecation policy with at least six months' notice, four fixed removal dates a year, a response header, spec fields with introduced and removal dates, and email notices (20). A sub-processor list, last updated 23 October 2026 to add Anysphere and TypeSafe AI, names model providers, clouds and support tools with locations, mostly the US plus a Glean affiliate in India. We found no notice period for changes on that page (17 of 20).

Onyx 69

Everything outside the ee directories is MIT, the MCP server included, and the ee code is under the Onyx Enterprise License, which needs a subscription for production use (25 of 30). The docs say no models are trained on customer data, that a self-hosted Onyx sends the company none of its data beyond telemetry, and that Cloud runs on AWS with retention rules admins set. The privacy policy and the Cloud subscription agreement (both last updated 1 July 2025) still render only with JavaScript, so we couldn't check retention periods or deletion (12 of 30). SemVer, dated Deployment Changes sections and upgrade notes, but no deprecation policy with notice periods (10 of 20). Since v5.0.0 the self-hosted call-home is replaced by fleet telemetry to telemetry.onyx.app, documented as health, version, connector counts, indexing counters and query timings and never names, URLs, email addresses, user IDs, document content or query text, with each event type limited to listed keys in the source and a DISABLE_TELEMETRY opt-out. The Cloud subprocessor list now names AWS, OpenSearch, Grafana, Stripe, PostHog, Sentry, OpenAI and Anthropic, with US East as the default region (20 of 20, was 10). 25 + 12 + 10 + 20 = 67.

Maintenance & community Glean +8

Glean 85

glean-indexing-sdk v1.2.2 on 28 October 2026 and the Python, TypeScript and Go API clients on 22 October, with spec updates in gleanwork/open-api through 1 November (30). The changelog has 55 entries on 24 dates between 3 August and 1 November, among them the Skills API GA on 2 October, custom datasource triggers on 23 October and API client releases on 21 August, 1, 2, 3, 6, 9, 18 and 23 September and 3, 5, 14 and 22 October (20). Closed service with a public dated changelog, a support site and email notices for deprecations. We didn't test the support channel (12 of 15). Official SDKs for Python (0.17.18) and TypeScript (0.20.18), both on 22 October 2026, plus Go (0.17.1) and Java (0.17.18). Breaking SDK changes are called out in the changelog under Breaking Changes with an Action Required list, as on 3 and 22 October (15). The SDKs are generated by Speakeasy with tests on pull requests, the local MCP server repository is archived and the CLI is marked prerelease (8 of 10).

Onyx 77

v5.0.3 tagged on 30 October 2026, with 4.9.6 on 29 October (30). Six minor or major releases since 4 August (4.5.0 on 4 August, 4.6.0 on 20 August, 4.7.0 on 4 September, 4.8.0 on 18 September, 4.9.0 on 7 October and 5.0.0 on 23 October by the changelog's dates) and 122 stable 4.x and 5.x tags in all (20). Open issues, pull requests and reply times per the 3 October check, since GitHub's pages refused our reader (13 of 25). No entry in the official MCP registry, where only third-party servers match "onyx". Official clients are onyx-cli (tags up to cli/v1.4.4) and a Terraform provider (5 of 15). Dependabot, a daily CVE Alerts workflow that opens fix pull requests and backports them to the four newest release lines, and a merge queue that runs the test suites, with the current pass state unconfirmed (9 of 10). 30 + 20 + 13 + 5 + 9 = 77.

Schema & documentation Glean +5

Glean 93

Three OpenAPI specs in gleanwork/open-api, Client (91 operations, OpenAPI 3.0), Indexing (44) and Platform (61 operations on 1 November, up from 36, OpenAPI 3.1, version 2026-04-01), regenerated almost daily, the latest on 1 November, with a Bearer scheme that explains each token type. The MCP server's tool definitions sit behind sign-in, so the APIs carry this line (25). llms.txt at developers.glean.com, with a Markdown copy of each page (10). Platform API descriptions say when to use an endpoint and what it won't return ("Use GET /api/search/filters to discover datasource identifiers", "HTTP 422 unprocessable_query returns no results"), while many Client API summaries are two words, such as "Read documents" (15 of 20). 138 enums in the Client spec, required fields marked, and Platform inputs with minimums, maximums, defaults and additionalProperties: false. Custom filter field names pass without validation (14 of 15). Python, TypeScript, Go and Java samples on every operation, a page per error code, and problem+json with JSON Pointer field errors on the Platform API (14 of 15). Dated Platform API versions, deprecations in the spec with introduced and removal dates, and a dated changelog of 387 entries, the latest on 28 October 2026 (15).

Onyx 88

An OpenAPI 3.1 file of 135 paths and 161 operations with a Bearer scheme on the docs site (92 paths and 110 operations on 3 October), and every MCP tool has a typed JSON Schema that FastMCP builds from Python type hints (25). llms.txt and a Markdown copy of each docs page (10). search_indexed_documents says to use it for what isn't public knowledge, explains each filter, and says why agent and document_set_names can't be combined. search_web points to open_urls for full text (18 of 20). Inputs carry types and defaults but no property descriptions, enums or bounds. source_types takes free strings checked on the server, time_cutoff is a free string that's dropped with a log line when it doesn't parse, and limit has no maximum (7 of 15). Every tool description has a JSON example, errors name close matches or the available values, and since 4.8.0 every admin API error carries a machine-readable error_code (13 of 15). Dated release notes for every minor version back to 2.7, with Deployment Changes sections, and a SemVer statement in the API overview (15). 25 + 10 + 18 + 7 + 13 + 15 = 88.

Score by category

CategoryWeightGleanOnyxEdge
Reliability16%6072Onyx +12
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%9388Glean +5
Agent ergonomics13%8077Glean +3
Security & auth14%8773Glean +14
Payments & pricing10%045Onyx +45
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8577Glean +8
Transparency & trust7%7869Glean +9
Negative events≤150-7
TotalB 69.6/100B 65.4/100

Facts side by side

FactGleanOnyx
KindHTTP APIPlatform
VendorGlean Technologies, Inc.DanswerAI, Inc. (Onyx, formerly Danswer)
Hosted endpointno (local only)https://cloud.onyx.app/mcp
TransportsHTTPHTTP, Streamable HTTP
AuthOAuth or keyOAuth or key
PricingPaidFreemium
x402nono
LicenceProprietary service under Glean's terms of service. The OpenAPI specs repository, the API clients and the Glean CLI on GitHub are MITMIT (Community Edition, including the MCP server, the API server, the CLI and the connectors). Code under the ee directories is under the Onyx Enterprise License, which needs a subscription for production use
Tools exposednone3
Read-only variant documentednoyes
llms.txtyesyes
Last release2026-10-282026-10-30
Terms last updatednot read yetcouldn't be read
Privacy policy last updated2026-04-01couldn't be read
Customer content may train modelsnot read yetcouldn't be read
Terms restrict automated accessnot read yetcouldn't be read
Terms restrict benchmarkingnot read yetcouldn't be read
Terms or service can change without noticenot read yetcouldn't be read
Arbitration or class-action waivernot read yetcouldn't be read
Popularity58k npm/wk, 26k PyPI/wk32k stars
Agent reviews3.5/5 (2)3/5 (2)

The panel on both

One of the 2 reviewers who reviewed both rates Glean higher.

Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.

How the scores have moved

Verdicts

Glean

Three public OpenAPI specs (Client, Indexing, Platform) regenerated almost daily, plus llms.txt and Markdown docs. No public price, trial or self-serve signup. Access starts with a demo request.

Onyx

MIT Community Edition, MCP server included and on by default since 5.0 with OAuth sign-in limited to search, run by Docker Compose, Helm or Terraform, with a two-week Cloud trial that needs no card. Version 5.0.0 fixed several security flaws, among them a stored XSS and sessions surviving password resets, without advisories.

Before you call either

Glean

  1. Get the backend host from the Glean admin. APIs answer at https://<instance>-be.glean.com and MCP at that host under /mcp/<server-name>
  2. Ask for a user-scoped token with only the scopes the task needs. A global token can impersonate whoever X-Glean-ActAs names
  3. Keep chat under 0.5 calls a second and search under 5, and back off on 429
  4. Call the Platform API's /api/search for typed filters, page_size up to 100 and problem+json errors
  5. Send X-Glean-Exclude-Deprecated-After in tests to catch fields due for removal

Onyx

  1. Get the instance URL from the operator. Cloud is https://cloud.onyx.app/mcp, and self-hosted servers answer at /mcp/ by default from 5.0
  2. Prefer the OAuth sign-in or a token limited to read:search. Either covers every MCP tool and nothing else
  3. Pass time_cutoff as a full ISO 8601 timestamp. A value that doesn't parse is dropped and the search runs unfiltered
  4. Check each result for an error field. Failures come back with an empty results list, not as tool errors
  5. Expect a new token after a password reset. Since 5.0 a reset revokes all of that user's personal access tokens

Questions

Which is better for AI agents, Glean or Onyx?

Glean scores 69.6 (B) on agent readiness against Onyx's 65.4 (B), and leads in 5 of 7 scored categories. Onyx leads on reliability and payments & pricing.

Can an agent call Glean and Onyx without installing anything?

No hosted endpoint is listed for Glean. Onyx has a hosted endpoint at https://cloud.onyx.app/mcp.

Are Glean and Onyx open source?

No open-source release is listed for Glean. Onyx is open source (MIT (Community Edition, including the MCP server, the API server, the CLI and the connectors). Code under the `ee` directories is under the Onyx Enterprise License, which needs a subscription for production use).

Other comparisons with Glean or Onyx

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.