Spider
- Free allowance
- Keyless /scrape: no key and no account, 4 requests a minute (free_rpm 4 in the API's free_tier_rate_limited answer, RateLimit-Limit 4 on keyless answers that pass); the documented keyless call is POST with a JSON body
- Verdict
- caveatusable, with a named limit or an unprobed read path
- Cost
- free tier
- Identity
- anonymousno card, no phone
- Binding limit
- request-ratekeyless /scrape: 4 requests a minute, named as free_rpm 4 beside paid_rpm 10000 in the API's free_tier_rate_limited answer and sent as RateLimit-Limit 4 on keyless answers that pass, the header the docs define as requests allowed a...
- Evidence
- measured33 sources; a live request was made and its output kept
- Freshness
- stale on 2027-01-02checked 2026-11-02 (live); stale on 2027-01-02
| Free tier | keyless /scrape: no key and no account, 4 requests a minute (free_rpm 4 in the API's free_tier_rate_limited answer, RateLimit-Limit 4 on keyless answers that pass); the documented keyless call is POST with a JSON body unit requests; metered on request-count; an account that passes the vendor's signup checks gets 2,500 credits worth USD 0.25 once, with no card; the balance does not expire; pay as you go is USD 1 a GB of bandwidth plus USD 0.0001 a CPU minute |
|---|---|
| Price | none stated |
| Limits |
keyless GETs of /scrape with a url query answer 400 with a zero-byte body and RateLimit-Limit 4, with RateLimit-Remaining 2 or 3 in a burst, until the throttle answers 429 with RateLimit-Limit 2 and RateLimit-Remaining 0, so RateLimit-Limit differs between the 400 and 429 answers; the docs answer a GET on a POST route with 400 and define RateLimit-Limit as requests allowed a minute |
| Binding | request-rate the keyless RateLimit-Limit of 4 binds before any volume; past it, a key or a per-request payment is needed |
| Identity | anonymous /scrape takes no key and no account; keyless /crawl answers Payment Required with an offer to pay per request in USDC on Base with no account, key or signup; an account takes an email and password, or Google, GitHub or Discord |
| Egress | not applicable: Markdown or JSON answers |
| Interface | rest REST /scrape |
| Tooling | REST over HTTPS with JSON bodies and a Bearer key, or no key on /scrape; official clients spider-client on PyPI and @spider-cloud/spider-client on npm, with Rust, Go and CLI clients in spider-rs/spider-clients; a hosted MCP server at mcp.spider.cloud; per-request USDC payment with an EVM wallet and an x402 client |
| Composes with | One page per keyless call through POST /scrape with a JSON body, for a first call or a connectivity check; a GET with a url query answers 400, and crawls, searches and sustained use move to a key or a per-request USDC payment |
| Reliability | keyless /scrape throttles with 429: a burst of 12 parallel GETs answered 400 eight times and 429 four times, each 429 carrying Retry-After, RateLimit-Limit 2, RateLimit-Remaining 0 and a 328 byte JSON body; Retry-After read 1, 3 and 10 seconds in separate runs, and a call made after a Retry-After of 3 had elapsed answered 400 again; measured with GET, keyless POST /scrape is documented and not measured |
| Notes | A keyless GET of /scrape with a url query answers 400 with a zero-byte body, the documented answer to a GET on a POST route, so it returns no page content; it does reach the keyless rate limiter, which answers a burst with 429. Keyless POST /scrape with a JSON body is documented and not measured. |
| Evidence note | measured liveness, the keyless RateLimit-Limit and the keyless 429 answer are measured with GET; the keyless POST /scrape, the account credits and the paid rates are documented, not measured |
| Parsed numbers |
|
Evidence 33
| date | source | what it proved |
|---|---|---|
| 2026-10-22 | api.spider.cloud/scrape?url=https://example.com | 400 with a zero byte body: the keyless route refused the request and sent no reason |
| 2026-10-30 | spider.cloud/register/ | a new account gets 2,500 credits worth USD 0.25 with no card, subject to signup checks |
| 2026-10-30 | probe: https://api.spider.cloud/scrape?url=https://example.com | GET /scrape?url=... with no key answered 400, 0 bytes, with RateLimit-Limit 4 and RateLimit-Reset 7 headers |
| 2026-10-30 | spider.cloud/register/ | the sign-up form takes an email and password, or Google, GitHub or Discord |
| 2026-10-30 | spider.cloud/faq/ | a credit balance does not expire |
| 2026-10-30 | spider.cloud/ | pay as you go bills USD 1 a GB plus USD 0.0001 a CPU minute |
| 2026-10-30 | spider.cloud/llms.txt | POST /scrape is documented as working with no key and no account |
| 2026-10-30 | probe: https://api.spider.cloud/scrape?url=https://example.com | a keyless GET of /scrape with a url query answered 400, 0 bytes, with RateLimit-Limit 4, RateLimit-Remaining 3 and RateLimit-Reset 52 and no payment demand: the API is live and admits a keyless /scrape request |
| 2026-10-30 | probe: https://api.spider.cloud/crawl | a keyless GET of /crawl answered with an x402 offer to pay per request in USDC on Base with no account, key or signup, priced USD 0.005 for up to 5 MB |
| 2026-10-30 | spider.cloud/docs/core/rate-limits/ | an account starts at 10,000 core API requests a minute and rises through four levels, the fourth at 80,000 |
| 2026-10-30 | spider.cloud/docs/overview/ | keyless runs from the documentation playground need no account and are capped at 25 a day |
| 2026-10-30 | probe: https://pypi.org/pypi/spider_client/json | PyPI serves spider-client 0.1.93, a Python SDK for the Spider Cloud API whose home page is spider-rs/spider-clients |
| 2026-10-30 | probe: https://api.spider.cloud/crawl | a keyless GET of /crawl answered 402 with 791 bytes of JSON and a PAYMENT-REQUIRED header |
| 2026-10-30 | probe: https://api.spider.cloud/data/credits | a keyless GET of /data/credits answered 401 with an Unauthorized JSON body, not the 402 the docs give for keyless routes other than /scrape |
| 2026-10-30 | probe: https://registry.npmjs.org/@spider-cloud%2Fspider-client/latest | npm serves @spider-cloud/spider-client 0.2.5 from the spider-rs/spider-clients repository |
| 2026-10-30 | spider.cloud/llms.txt | a GET to any content route answers 400, so a GET of /scrape tests the method and not the keyless route |
| 2026-10-30 | spider.cloud/llms-full.txt | the keyless tier is documented as a first call and connectivity check; sustained use needs a key |
| 2026-10-30 | spider.cloud/docs/core/rate-limits/ | the docs define the RateLimit-Limit header as requests allowed a minute |
| 2026-10-30 | spider.cloud/docs/integrations/x402/ | per-request payment in USDC on Base through x402 needs no account, key or signup |
| 2026-10-30 | spider.cloud/docs/libraries/ | official SDKs exist for Python, JavaScript, Rust and Go, and a CLI installs with cargo |
| 2026-11-01 | probe: https://api.spider.cloud/scrape?url=https://example.com | the keyless 429 answer is a JSON body with code free_tier_rate_limited that names free_rpm 4 and paid_rpm 10000: the free tier is 4 requests a minute and an account 10000 |
| 2026-11-02 | probe: https://api.spider.cloud/scrape?url=https://example.com | a burst of 12 parallel keyless GETs of /scrape answered 400 eight times and 429 four times; each 429 carried Retry-After 10, RateLimit-Limit 2 and RateLimit-Remaining 0, each 400 carried RateLimit-Limit 4 and RateLimit-Remaining 2 or 3 |
| 2026-11-01 | probe: https://api.spider.cloud/scrape?url=https://example.com | the keyless 429 answer tells the caller to slow down, or create an account and add credits for up to 10,000 requests/min |
| 2026-11-01 | probe: https://api.spider.cloud/scrape?url=https://example.com | a keyless GET of /scrape answered 429 with Retry-After 3, RateLimit-Limit 2, RateLimit-Remaining 0 and RateLimit-Reset 3 |
| 2026-11-01 | probe: https://api.spider.cloud/scrape?url=https://example.com | a keyless GET of /scrape answered 429 with Retry-After 1, RateLimit-Limit 2, RateLimit-Remaining 0 and RateLimit-Reset 1 |
| 2026-11-01 | probe: https://api.spider.cloud/scrape?url=https://example.com | a keyless GET probed at 19:27:59.949Z, 3.2 seconds after a 429 with Retry-After 3 probed at 19:27:56.708Z, answered 400 with RateLimit-Limit 4, RateLimit-Remaining 3 and RateLimit-Reset 59 |
| 2026-11-01 | spider.cloud/scrapers/indeed-scraper/ | the vendor's scraper pages label the live keyless run with a rate of 4 a minute |
| 2026-11-01 | spider.cloud/agent-skill/SKILL.md | the docs say going over the keyless limit returns 429 and tell the caller to honor Retry-After |
| 2026-11-01 | spider.cloud/docs/core/rate-limits/ | an account starts at 10,000 core API requests a minute and rises through four levels, the fourth at 80,000 |
| 2026-11-01 | spider.cloud/docs/overview/ | keyless runs from the documentation playground need no account and are capped at 25 a day |
| 2026-11-01 | spider.cloud/llms-full.txt | POST /scrape is documented as working with no Authorization header and no account |
| 2026-11-01 | spider.cloud/llms-full.txt | the docs say /scrape answers with a JSON array holding one object for its single page, with a content field |
| 2026-11-01 | spider.cloud/llms-full.txt | the docs call the keyless tier a way to make a first call and confirm connectivity, and say anything sustained needs a key |
Unknown values 1
- per object cap stated: the field does not apply to this kind of offer