# AX Check: curl.se Checked 2026-09-18. curl.se offers agents no machine-readable docs or index to work from Homepage has no Markdown option and llms.txt returns a 404, so agents get no documentation index; 21 of 23 checklist items were left unassessed as a result. ## Onboarding needs a login ## Coding sessions All three independent sessions completed, but without pricing or product info on the site, they diverged wildly: one refused to state any pricing, one invented detailed per-token API pricing for an unrelated hypothetical product, and one described a free-tier API rate-limit structure. None of these reflect curl's actual nature as free open-source software, showing the agents had no reliable source to ground their answers. ### DeepSeek V4 Pro [View public run](https://agents.withgauge.com/p/runs/e46df317-a2aa-4d01-b7e4-ef78caad9ab3) · [Read transcript](https://www.ax-check.com/curl.se/sessions/deepseek.json) Final output explicitly refuses to state any pricing ('I won't invent a fake product or a guessed pricing model') — no figure or assumptions given at all. #### End-to-end onboarding - **End-to-end onboarding not demonstrated**: No product, SDK, API endpoint, or credentials existed anywhere in the sandbox. The repo contained only a .gitignore and a single bootstrap commit, and the agent's own auth/config files were empty. No product operation was attempted or possible. Event 20: ```text ---ALL FILES INCL HIDDEN--- ./.gitignore ---ANY MD/TXT ANYWHERE NON-GIT--- /sandbox/.pi/agent/auth.json /sandbox/.pi/agent/models-store.json /sandbox/.pi/agent/models.json ``` Event 25: ```text --- auth.json (keys redacted) --- {}--- models-store.json sample --- {} ``` Event 27: ```text There is no product in this workspace for me to act on. ``` #### Hallucinated URLs None identified in this transcript. #### Blockers - **No product artifacts present in the workspace**: The task asked for a curl example, pricing summary, and manageability assessment for a hosted product, but the repo held only a .gitignore and an initial bootstrap commit — no README, manifest, SDK, API base URL, or docs. This is a test-environment setup issue rather than a product defect or agent error; the agent correctly identified the gap instead of fabricating a workflow. Event 5: ```text total 16 drwxr-xr-x 3 root root 4096 Sep 18 08:00 . drwxr-xr-x 7 root root 4096 Sep 18 20:15 .. drwxr-xr-x 8 root root 4096 Sep 18 08:00 .git -rw-r--r-- 1 root root 408 Sep 18 08:00 .gitignore ``` Event 20: ```text ---ANY MD/TXT ANYWHERE NON-GIT--- /sandbox/.pi/agent/auth.json /sandbox/.pi/agent/models-store.json /sandbox/.pi/agent/models.json ``` Event 27: ```text So there is no developer workflow I can act on: no endpoint to curl, no SDK to call, no pricing page or plan details to read. I won't invent a fake product or a guessed pricing model. ``` #### Suggested Changes None identified in this transcript. ### Kimi K3 [View public run](https://agents.withgauge.com/p/runs/64e93daa-1a6c-4f4f-967d-90b922b7b9b7) · [Read transcript](https://www.ax-check.com/curl.se/sessions/kimi.json) Final output gives per-token pricing (gpt-4o-mini ~$0.15/M in, $0.60/M out; gpt-4o ~$2.50/M in, $10/M out) and names assumptions: pay-as-you-go model, no subscription, batch API discount, prices 'subject to change'. #### End-to-end onboarding - **Onboarding needs a login**: The agent found an empty repo with no product identified, then defaulted to testing the OpenAI API on its own initiative. It hit a 401 Unauthorized on an unauthenticated request and confirmed no API key existed in the environment. It explicitly stopped and asked the human to supply an OpenAI account/API key, never obtaining credentials itself or completing any authenticated call. Event 20: ```text HTTP 401 --- { "error": { "message": "Missing bearer authentication in header", "type": "invalid_request_error", "param": null, "code": null } } ``` Event 24: ```text no OpenAI-related env vars set ``` Event 26: ```text There's no anonymous sandbox or test mode, and no key in this environment. Since you don't have an account and I won't fake credentials or mock it locally, I can't execute a real example. ``` #### Hallucinated URLs None identified in this transcript. #### Blockers - **No product specified in the empty repo**: The initial workspace was an empty git repo with only a .gitignore, giving the agent no indication of which product to test. This is a test-environment/setup issue rather than a product defect; the agent correctly flagged it and asked for clarification instead of guessing blindly. Event 8: ```text 9ebf5fb Initial workspace * main # Keep generated and vendored files out of source diffs. # Logs / OS cruft *.log .DS_Store # JavaScript / TypeScript node_modules/ .next/ dist/ ``` Event 10: ```text There's no product to work with. The working directory (`/sandbox/repo`) contains only an empty git repo with a `.gitignore` — no code, no README, no mention of any product, API, or SDK. ``` - **Missing OpenAI API key blocks any real execution**: After the agent unilaterally chose to test the OpenAI API, it found no credentials in the sandbox environment and no anonymous/test mode on the API itself. This is a normal authentication requirement (product behavior), not a product defect, but it is an unresolved blocker for this session since no human supplied a key and the agent could not self-provision one. Event 20: ```text HTTP 401 --- { "error": { "message": "Missing bearer authentication in header", "type": "invalid_request_error", "param": null, "code": null } } ``` Event 24: ```text no OpenAI-related env vars set ``` Event 26: ```text I can't execute a real example. ``` #### Suggested Changes None identified in this transcript. ### Qwen 3.8 Max [View public run](https://agents.withgauge.com/p/runs/dc2c5505-5a0c-458f-89a0-9756917ef870) · [Read transcript](https://www.ax-check.com/curl.se/sessions/qwen.json) Final output states the API is free with the assumption basis spelled out: 60 req/hour unauthenticated vs 5,000 req/hour with a free token, and higher paid tiers for heavier workload. #### End-to-end onboarding - **End-to-end onboarding not demonstrated**: The agent never engaged a real product's authenticated developer workflow. It first found the sandboxed repo empty and correctly stopped rather than guessing a product. On retry, it explicitly ruled out Stripe because it would require a signup/API key, and substituted the free, keyless GitHub REST API instead. All calls made (rate_limit, repos/stripe/stripe-node) were unauthenticated read-only GET requests, so no credentials were ever obtained or used, and no authenticated operation (create/write) was demonstrated on any product. Event 25: ```text Stripe's API can't be called without an account — even test mode requires a secret key from a signup, so it fails your "no key" constraint. So I picked the GitHub REST API ``` Event 22: ```text curl -s https://api.github.com/repos/stripe/stripe-node ``` #### Hallucinated URLs None identified in this transcript. #### Blockers - **No product identified in the sandbox repo**: The sandbox repo the agent was given contained no product code, SDK, API spec, or docs — just an empty git repo. This is a test-environment limitation, not a product defect: the agent correctly recognized it had no subject to build a curl example against and stopped rather than inventing one. Event 10: ```text I checked the environment first: `/sandbox/repo` is empty (only a `.git` directory and `.gitignore`) — there's no product, SDK, API reference, quickstart, or docs here to act on. ``` - **Stripe requires signup/API key, so it was dropped as the example target**: When the agent picked Stripe as a candidate product to demo, it recognized that even test-mode calls require a secret key obtained via signup. Since the session constraints ruled out needing new credentials, the agent switched targets to GitHub's key-free API rather than resolving the credential need. This reflects normal product authentication requirements, not a product defect. Event 25: ```text Stripe's API can't be called without an account — even test mode requires a secret key from a signup, so it fails your "no key" constraint. ``` #### Suggested Changes None identified in this transcript. ### Task given to each agent Help me build a simple example using curl. 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: F · 0/100 (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. ``` - **Failed** — llms.txt provides an actionable documentation index ```text https://curl.se/llms.txt returns HTTP 404, so no documentation index exists. ``` - **Skipped** — llms.txt provides navigation guidance ```text llms.txt is 404, so no navigation guidance content is available to assess. ``` - **Skipped** — llms.txt mentions offered API, MCP, and skills ```text llms.txt is 404, so no API/MCP/skills mentions are available to assess. ``` - **Skipped** — A compact guide representation exists ```text No site-published Markdown or compact guide representation was fetched; only HTML pages. ``` - **Skipped** — A focused guide is directly retrievable ```text No focused guide representation was directly retrieved in the fetched evidence. ``` - **Skipped** — Equivalent instructions fit a token budget ```text No compact guide was fetched, so token budget cannot be measured. ``` - **Skipped** — Product-docs links survive format changes ```text Homepage Markdown is unsupported, so link preservation across formats is unassessed. ``` - **Skipped** — The compact guide is independently actionable ```text No compact agent guide was fetched; only homepage and a 404 llms.txt were supplied. ``` - **Skipped** — Install and next-step links resolve ```text No install or next-step pages were fetched, so link resolution cannot be judged. ``` ### Onboarding - **Skipped** — Docs lead to a relevant quickstart ```text No fetched docs or quickstart pages were supplied; only a 404 llms.txt and a YouTube page. ``` - **Skipped** — Installation commands are extractable ```text No installation or download page content was fetched to extract commands from. ``` - **Skipped** — Code examples are available without interaction ```text No documentation or example pages were fetched; code examples cannot be judged. ``` - **Skipped** — Prerequisites and auth boundaries are explicit ```text No quickstart or API page fetched; prerequisites and auth boundaries not observable. ``` ### Pricing - **Skipped** — Pricing is readable without interaction ```text Not applicable: the relevant product is free. curl is free open source software; no paid developer product pricing applies. ``` - **Skipped** — Prices are stated, not gated ```text Not applicable: the relevant product is free. curl is free open source software; no paid developer product pricing applies. ``` - **Skipped** — Pricing units and limits are explicit ```text Not applicable: the relevant product is free. curl is free open source software; no paid developer product pricing applies. ``` - **Skipped** — Agents identify pricing and its assumptions ```text Not applicable: the relevant product is free. curl is free open source software; no paid developer product pricing applies. ``` ### Activation - **Skipped** — An API reference or OpenAPI spec is reachable ```text No API reference or OpenAPI spec was fetched; only a 404 llms.txt and a YouTube page. ``` - **Skipped** — An MCP server is documented and well-formed ```text No MCP server documentation was offered or fetched in the supplied evidence. ``` - **Skipped** — A CLI install path is documented ```text No CLI install page was fetched; the 404 llms.txt and YouTube page show no install path. ``` - **Skipped** — SDK packages resolve on their registries ```text No SDK package registry lookup was supplied, so package resolution cannot be judged. ``` - **Skipped** — Agent skills are published ```text No agent skills documentation was offered or fetched in the supplied evidence. ``` [Full report data](https://www.ax-check.com/curl.se/report.json)