Hosted routers and marketplaces - verified
Requesty
Requesty is listed under hosted routers and marketplaces. Its first listed fee, in its own words: 5% markup, on model cost (pay as you go). It states OpenAI-compatible and Anthropic-compatible requests. Routing stated: fallbacks and load balancing. Read from the vendor's own pages on October 18, 2026.
Read from docs.requesty.ai, requesty.ai. Docs - Pricing - Compare every gateway
Status
- Status, as we read it
GA (self-serve)inferred
Our reading: The page implies this status without labelling it.
“No per-seat pricing, no minimum spend.”
requesty.ai
How Requesty charges
Every fee line Requesty states, in its own words and units, with the quote. Nothing is converted.
| Product | Fee, in the vendor's words | Applies to | Quote and source |
|---|---|---|---|
| Requesty | 5% markup | on model cost (pay as you go) | “You pay for what your application spends, plus 5%.” |
| $0 | Free plan (free models, 200 requests per day) | “The full platform, on free models” | |
| Custom | Enterprise plan | “For teams with advanced needs” |
Access and keys
- Free tier
free models, 200 requests per day, no credit card
“Start free with 200 requests a day, or pay as you go for 5% on model cost.”
requesty.ai - Bring your own key
yes (pay-as-you-go plan)
“Bring your own keys”
requesty.ai - BYOK fee
- Not stated
Whether the 5% markup applies to BYOK traffic is not answered in the captured page text (FAQ answers were collapsed).
API formats
- OpenAI-compatible
yes
“Requesty returns an OpenAI-compatible response.”
docs.requesty.ai - Anthropic-compatible
yes (Anthropic SDK against /anthropic/v1/messages)
“Same for the Anthropic SDK against /anthropic/v1/messages.”
docs.requesty.ai - OpenAI base URL
https://router.requesty.ai/v1
“base_url="https://router.requesty.ai/v1"”
docs.requesty.ai - Anthropic base URL
- Not stated
Path stated as /anthropic/v1/messages; full base URL for the Anthropic SDK not stated on the checked page.
Routing
- Fallbacks
yes (fallback policies)
“If the primary model errors or times out, Requesty tries the next one.”
docs.requesty.ai - Load balancing
yes (load balancing and latency routing)
“Fallbacks, load balancing, and latency routing across 30+ providers.”
docs.requesty.ai - Automatic model routing
- Not stated
Automatic provider selection for bare model names is documented; automatic model selection is not stated.
Caching
- Exact-match cache
- Not stated
Cache type not stated on the checked pages.
- Semantic cache
- Not stated
Not stated in the checked pages.
- Caching, as listed
auto-caching on repeated prompts
“Cut costs up to 80% on repeated prompts with zero configuration.”
docs.requesty.ai
Controls and observability
- Spend controls
spend limits and budget caps (pay as you go)
“Spend limits & budget caps”
requesty.ai - Rate limits
- Not stated
Not stated beyond the free-plan request limit.
- Guardrails
guardrails and PII detection (Enterprise)
“Guardrails & PII detection”
requesty.ai - Observability
spend tracking and analytics; advanced observability on pay as you go
“Spend tracking & analytics”
requesty.ai - Telemetry and log export
"Data exports" listed in plan comparison (tier not determinable from captured table)
“Data exports”
requesty.ai
Data, deployment and license
- Data logging and retention
- Not stated
No retention or ZDR statement in the checked pages.
- Regions
EU data residency on every plan
“Routing, caching and EU data residency are included on every plan”
requesty.ai - Self-hosting
- Not stated
Hosted service; not stated.
- License
- Not stated
Hosted proprietary service; no open-source license stated.
Vendor-stated counts
Model and provider counts as Requesty states them, dated. When two pages disagree, both are shown.
600+ models (vendor-stated, Oct 18 2026)
“All 600+ models”
requesty.ai 20+ providers (pricing comparison table) (vendor-stated, Oct 18 2026)
“20+ providers”
requesty.ai 30+ providers (docs quickstart) (vendor-stated, Oct 18 2026)
“across 30+ providers”
docs.requesty.ai
Research notes
Provider counts differ between the pricing page (20+) and the docs (30+); both are recorded.
FAQ
How much does Requesty cost?
In Requesty's own terms: 5% markup (on model cost (pay as you go)); $0 (Free plan (free models, 200 requests per day)); Custom (Enterprise plan). As of October 18, 2026; fees are not converted.
Does Requesty support bring your own key?
Yes (pay-as-you-go plan), per the vendor: "Bring your own keys". BYOK fee: not stated. Whether the 5% markup applies to BYOK traffic is not answered in the captured page text (FAQ answers were collapsed).
Is Requesty OpenAI-compatible or Anthropic-compatible?
OpenAI: yes, base URL https://router.requesty.ai/v1. Anthropic: yes (Anthropic SDK against /anthropic/v1/messages).
Sources
- Source: docs.requesty.ai/quickstart
- Source: requesty.ai/pricing
Last verified from each vendor's own docs, pricing pages, READMEs and license files; every quote was string-matched against the page that day. Fees are shown in the vendor's own words and change without notice, so check the linked source before you commit.
Newsletter
Know when a gateway changes its fees
One email per week. New model releases, routing and pricing changes and what they mean for what you build.