# AX Check: glama.ai Checked 2026-10-29. Docs are clear, but pricing stays hidden behind sign-up A compact guide links the API, directory, and MCP docs cleanly, but no pricing page was ever fetched — rates only surfaced via live API lookups or an archived page. ## Onboarding needs a login ## Coding sessions All three sessions completed. One found no dollar figures since the rate card sits behind sign-up, naming the free and metered-hosting plan basis instead. Another pulled specific tier prices ($26, $80) from an archived page. The third quoted live per-token API rates while flagging that full billing details require login. ### DeepSeek V4.1 Flash [View public run](https://agents.withgauge.com/p/runs/1100dbb9-4736-4b2d-892b-3644439592ae) · [Read transcript](https://www.ax-check.com/glama.ai/sessions/deepseek.json) Gave no dollar figure because the rate card is behind sign-up. Named the basis: free account, workspace plan, hosting metered by machine hours and outbound data (seq 23, seq 106, seq 129) #### End-to-end onboarding - **End-to-end onboarding not demonstrated**: Agent never created a Glama account or obtained an API key, so no authenticated product operation was demonstrated. The only live call that returned real data (POST /v1/telemetry/usage) requires no credentials at all, and the bearer-auth endpoints (GET /v1/servers, /v1/connectors) were only called once, unauthenticated, returning 401. The example script's authenticated branch was left unexercised with a 'skipped' message. Event 105: ```text === telemetry with real record === {"success":true} HTTP 200 ``` Event 43: ```text {"error":{"code":"unauthorized","message":"This endpoint requires an API key. Create one at https://glama.ai/settings/api-keys.\n\nUse of this data is governed by the API Data License ``` Event 127: ```text [servers] skipped: set GLAMA_API_KEY to run this call. ``` #### Hallucinated URLs None identified in this transcript. #### Blockers - **Authenticated registry endpoints require a manually-created API key**: GET /v1/servers and /v1/connectors both demand a bearer token obtainable only by signing up and visiting the settings page. This is normal product auth, not a defect, and the agent correctly avoided creating an account since the task didn't call for human sign-up. It left the authenticated code path unexercised rather than treating it as a hard stop. Event 43: ```text This endpoint requires an API key. Create one at https://glama.ai/settings/api-keys. ``` - **Public pricing page redirects to sign-up with no rate card**: Fetching glama.ai/pricing returned a sign-up form instead of dollar figures. The FAQ/hosting pages only describe the billing model (workspace plan plus metered machine hours and egress) and explicitly state the real rate card lives behind login in subscription settings. This is a product content gap for prospective users evaluating cost upfront, not an agent error. Event 23: ```text Sign Up | Glama All-in-one AI workspace ``` Event 106: ```text The current rate card is in your workspace under subscription settings . ``` #### Suggested Changes - **Publish an actual rate card on the /pricing page**: glama.ai/pricing currently serves only a sign-up form with no dollar amounts; the FAQ states real costs are hidden in workspace subscription settings post-login. Add at least illustrative pricing tiers or a machine-hour/egress rate table to the public /pricing page so prospective developers can estimate cost before creating an account. Verify by fetching https://glama.ai/pricing unauthenticated and confirming numeric rates appear in the page text. Event 23: ```text Sign Up | Glama All-in-one AI workspace A ChatGPT-like chat experience for talking with AI. ``` Event 106: ```text How much does hosting cost? Hosting is billed on top of your workspace plan; infrastructure cost is metered by machine hours and egress. ``` - **Link an official SDK or client library from the MCP reference page**: No official Glama SDK package exists on npm; the only match ('glama') is an unrelated third-party framework for building MCP servers, and '@glama/sdk' 404s. Add a link to a maintained client library (or explicit confirmation that raw REST/fetch is the intended integration path) on https://glama.ai/mcp/reference so developers stop searching npm for a nonexistent official package. Event 55: ```text npm error 404 Not Found - GET https://registry.npmjs.org/@glama%2fsdk - Not found ``` Event 55: ```text glama 1.0.0 - A TypeScript framework for building MCP servers. ``` ### Kimi K3 [View public run](https://agents.withgauge.com/p/runs/2567b3ab-8657-465b-8724-948b40a94286) · [Read transcript](https://www.ax-check.com/glama.ai/sessions/kimi.json) Priced the free, $26 and $80 tiers from an archived page it identified as such, with add-on fees and token billing at provider prices with no markup (seq 42, seq 44) Session findings unavailable. ### Qwen 3.8 Max [View public run](https://agents.withgauge.com/p/runs/8b574307-73ad-4c4b-8152-f6749c3c3b67) · [Read transcript](https://www.ax-check.com/glama.ai/sessions/qwen.json) Quoted per-model rates from the live API (qwen3-30b-a3b $0.13/$0.52, OpenAI ~$1/$5 per 1M) as pay-as-you-go per token; noted billing details behind login (seq 43, seq 61, seq 68, glama_example/README.md) #### End-to-end onboarding - **Onboarding needs a login**: The agent found Glama's hosted OpenAI-compatible gateway and successfully called the unauthenticated /models endpoint, but the actual product operation (chat completion) requires an API key obtained only via a web sign-up flow. The agent explicitly could not complete this sign-up itself and stopped, never obtaining or using real credentials. Event 34: ```text {"error":{"code":"unauthorized","message":"Missing authorization header"}} ``` Event 68: ```text The main friction: one API key must be created through a web sign-up, which I can't do. ``` Event 61: ```text 305 models available on the Glama gateway ``` #### Hallucinated URLs - **Guessed /gateway path redirected harmlessly, not a true hallucination**: The agent tried https://glama.ai/gateway which 301-redirected to the real /mcp page rather than failing, so this does not meet the bar for a demonstrably invented-and-failed URL. Omitted from hallucination list per instructions since it did not fail. Event 20: ```text == https://glama.ai/gateway 404 https://glama.ai/docs == https://glama.ai/api 404 https://glama.ai/api == https://glama.ai/gateway 200 https://glama.ai/mcp ``` #### Blockers - **API key requires manual web sign-up**: Glama's chat completion endpoint (the core product operation) returns 401 without an API key, and keys can only be created through the hosted sign-up page, not self-service in this sandboxed session. This is a normal authentication requirement, not a product defect — the agent correctly identified it as the one blocked step rather than attempting a workaround. Event 35: ```text {"error":{"code":"unauthorized","message":"Missing authorization header"}} ``` Event 68: ```text Blocked on one thing: `chat.py` needs `GLAMA_API_KEY`. I verified the endpoint returns `401 Missing authorization header` without it. ``` - **No dedicated pricing or docs page resolves cleanly**: Both /pricing and /docs 404'd or redirected to sign-up/category pages instead of a clear pricing or developer-docs page, forcing the agent to reconstruct pricing info from the raw /models API response instead of a human-readable source. Event 20: ```text == https://glama.ai/docs 404 https://glama.ai/docs ``` Event 28: ```text 200 https://glama.ai/sign-up?returnPath=%2Fsettings%2Fworkspace%2Fsubscription Sign Up | Glama All-in-one AI workspace ``` Event 36: ```text 404 https://glama.ai/mcp/pricing ``` #### Suggested Changes - **Add a public pricing page at /pricing**: Requests to https://glama.ai/pricing and https://glama.ai/mcp/pricing both failed to show plan/rate info (one redirected to sign-up, the other 404'd), forcing reliance on the raw /models API JSON for token rates. Publish a plain pricing page and verify it loads without a redirect to sign-up by curling https://glama.ai/pricing directly. Event 28: ```text 200 https://glama.ai/sign-up?returnPath=%2Fsettings%2Fworkspace%2Fsubscription Sign Up | Glama All-in-one AI workspace ``` Event 36: ```text 404 https://glama.ai/mcp/pricing ``` - **Add a developer docs page at /docs**: A request to https://glama.ai/docs returned 404, leaving no canonical developer quickstart; the agent had to infer the gateway's OpenAI-compatible usage from the raw /models endpoint and trial-and-error curl calls. Publish API docs at /docs covering the gateway base URL and auth, then confirm with a direct GET to https://glama.ai/docs returning 200. Event 20: ```text == https://glama.ai/docs 404 https://glama.ai/docs ``` ### Task given to each agent Help me build a simple example using Glama. Tell me how pricing works, and briefly tell me whether this product will be easy for you to manage. Let me know if you get blocked. If this product has no developer workflow you can act on, say so plainly and stop. Stay light: use the hosted product through its SDK or API. Do not start local service stacks or wait for long-running commands; if the quickstart requires either, say so plainly and stop. No product credentials were supplied and no purchases were authorized. ## Score: B · 76/100 (provisional) Grades come from completed site checks. Coding sessions and skipped checks do not affect the score. ### Clarity - **Failed** — Homepage answers Markdown requests ```text Homepage returned text/html for a text/markdown request; no Markdown representation offered. ``` - **Pass** — llms.txt provides an actionable documentation index ```text llms.txt lists API reference, OpenAPI, directory, inspector, gateway, hosting and schema links. ``` - **Pass** — llms.txt provides navigation guidance ```text llms.txt groups links under Programmatic access, Directory, Working with MCP servers, and How the directory works. ``` - **Pass** — llms.txt mentions offered API, MCP, and skills ```text llms.txt names the MCP API, OpenAPI description, and API catalog; no skills offered. ``` - **Pass** — A compact guide representation exists ```text llms.txt is a compact Markdown guide to Glama's directory, API, and tools. ``` - **Pass** — A focused guide is directly retrievable ```text llms.txt fetched directly at glama.ai/llms.txt as text/plain Markdown. ``` - **Pass** — Equivalent instructions fit a token budget ```text llms.txt is 824 tokens, well under the 8000-token budget. ``` - **Skipped** — Product-docs links survive format changes ```text Homepage Markdown unsupported, so link preservation across formats cannot be measured. ``` - **Pass** — The compact guide is independently actionable ```text llms.txt gives a compact, organized index of Glama's API, directory, and MCP surfaces with starting guidance. ``` - **Pass** — Install and next-step links resolve ```text Fetched install/next-step links (API reference, OpenAPI, servers, connectors) all returned 200. ``` ### Onboarding - **Skipped** — Docs lead to a relevant quickstart ```text llms.txt links to MCP API reference and directory, but no fetched quickstart page. ``` - **Skipped** — Installation commands are extractable ```text No fetched Glama install or CLI page; only llms.txt index was retrieved. ``` - **Skipped** — Code examples are available without interaction ```text No fetched Glama docs page with code examples; only llms.txt index retrieved. ``` - **Skipped** — Prerequisites and auth boundaries are explicit ```text llms.txt names an MCP API but no fetched page states auth or prerequisites. ``` ### Pricing - **Skipped** — Pricing is readable without interaction ```text No Glama pricing page was fetched; only foreign products' pricing pages appear. ``` - **Skipped** — Prices are stated, not gated ```text No Glama pricing page was fetched; only foreign products' pricing pages appear. ``` - **Skipped** — Pricing units and limits are explicit ```text No Glama pricing page was fetched; only foreign products' pricing pages appear. ``` - **Pass** — Agents identify pricing and its assumptions ```text 3 of 3 sessions were judged on pricing; 0 fell short. DeepSeek V4.1 Flash: Gave no dollar figure because the rate card is behind sign-up. Named the basis: free account, workspace plan, hosting metered by machine hours and outbound data (seq 23, seq 106, seq 129) Kimi K3: Priced the free, $26 and $80 tiers from an archived page it identified as such, with add-on fees and token billing at provider prices with no markup (seq 42, seq 44) Qwen 3.8 Max: Quoted per-model rates from the live API (qwen3-30b-a3b $0.13/$0.52, OpenAI ~$1/$5 per 1M) as pay-as-you-go per token; noted billing details behind login (seq 43, seq 61, seq 68, glama_example/README.md) This behavioural item does not affect the fast grade. ``` ### Activation - **Pass** — An API reference or OpenAPI spec is reachable ```text Glama's MCP API reference and OpenAPI description are both linked and fetched successfully. ``` - **Pass** — An MCP server is documented and well-formed ```text Glama documents MCP servers and connectors with tools, schemas, and TDQS scoring. ``` - **Skipped** — A CLI install path is documented ```text No Glama CLI install path is documented in the fetched evidence. ``` - **Skipped** — SDK packages resolve on their registries ```text No Glama-owned SDK or CLI package registry lookup was supplied. ``` - **Skipped** — Agent skills are published ```text No Glama agent skills are published in the fetched evidence. ``` [Full report data](https://www.ax-check.com/glama.ai/report.json)