Requesty
6.8/10Requesty is a managed LLM gateway that gives development teams one OpenAI-compatible endpoint to hundreds of AI models, with routing, failover, caching and spend controls.
Data last updated November 2026 · Our methodology
Overview
Requesty is a hosted AI gateway run by London-based Requesty Ltd. Developers send requests to a single OpenAI-compatible endpoint, router.requesty.ai/v1, and reach more than 600 models from over 30 providers, including OpenAI, Anthropic, Google, Meta and AWS. Switching over usually means changing only the base URL and the API key. Routing policies are set in a dashboard and called by name, so teams can change fallback chains, percentage load balancing and latency-based routing without redeploying code. The gateway also offers provider-native prompt caching, an MCP gateway, guardrails for PII and prompt injection, and analytics that track tokens, cost, latency, cache hits and failovers. EU data residency is included on every plan, which matters for teams that must meet GDPR. Pricing has no seats and no minimum spend. The free plan allows 200 requests a day on free models. Pay as you go adds a 5 percent markup on model cost and supports bring-your-own-keys. Enterprise is priced on request and adds SSO, RBAC, audit logs, approved-model policies and custom SLAs. Requesty suits startups and SaaS teams that use several model providers and want failover and budget caps without running LiteLLM themselves. The main drawback is the 5 percent markup, which adds up at high monthly spend. Its provider network is also smaller than OpenRouter's, and its prompt management and evaluation tools are thinner than Portkey's.
Our Rating
AI Software6.9/10
Hosting & Infrastructure6.7/10
Who Requesty is best for
Best software for DevelopersGood fit
It works as a drop-in replacement for the OpenAI SDK, with routing policies configured in a dashboard.
Failover across providers and per-key budget caps protect AI features in production.
It has no seats or minimum spend, and the free tier needs no credit card.
Less suited
SSO, RBAC, audit logs and custom SLAs are available, but only on the custom-priced Enterprise plan.
Pros & Cons
Pros
- The endpoint is OpenAI-compatible, so migrating usually means changing only the base URL and API key.
- Fallback chains, weighted load balancing and latency routing are set in a dashboard and can be changed without redeploying.
- EU data residency is included even on the free tier.
- There are no per-seat fees or minimum spend, and bring-your-own-keys keeps existing provider rates.
- Analytics cover cost, latency, cache hits and failovers, with alerts sent to Slack and webhooks.
Cons
- The 5% markup on model cost becomes expensive at high monthly spend, compared with self-hosted LiteLLM.
- Its provider network is smaller than OpenRouter's (20 to 30 providers versus 70 or more).
- SSO, RBAC, PII controls and custom SLAs are only on the Enterprise plan.
- Prompt management and evaluation tools are thinner than Portkey's.
- The free tier is capped at 200 requests a day and covers free models only.
Key Features
Requesty Pricing
Free
- Free models only
- Community support
- EU data residency
- 200 requests per day
- No credit card required
- Spend tracking and analytics
- Routing, caching and fallbacks
Enterprise
- Approved-model policies
- Full RBAC and audit logs
- Service accounts for CI/CD
- Guardrails and PII detection
- Team and group spend controls
- Dedicated support and custom SLAs
- SSO (Okta, Azure AD, Google Workspace, OIDC)
Pay as you go
- 600+ models
- MCP Gateway
- Email support
- Bring your own keys
- Advanced observability
- Spend limits and budget caps
- Routing policies, caching and fallbacks
- 5% markup on model cost, no seat fees or minimum spend
Frequently Asked Questions
Pay as you go adds a 5% markup to the provider's model cost, so a model priced at $10 per million tokens costs $10.50. There are no seat fees or minimum spend.
Yes. The free plan gives access to free models, up to 200 requests a day, with no credit card required.
Yes. Bring-your-own-keys is included on the Pay as you go and Enterprise plans.
EU data residency is included on all plans, and a DPA is available. Requesty reports SOC 2 Type II compliance, according to a competitor's comparison article. Logging can be turned off entirely.
You point the OpenAI SDK at router.requesty.ai/v1 and name models in provider/model-name format. Streaming works as normal.
The free plan gets community support and Pay as you go gets email support. Enterprise gets dedicated support with custom SLAs.
Similar Tools
See all alternatives
Amazon's public cloud platform, with over 200 pay-as-you-go services for compute, storage, databases, networking and AI, used by startups through large enterprises and governments.
Portkey is an AI gateway and LLMOps control plane that routes, monitors, and governs requests to 1,600+ LLMs for engineering teams and enterprises.
Northflank is a Kubernetes-based developer platform for deploying services, jobs, databases and GPU workloads on its managed cloud or on customers' own cloud accounts.
Ranked near Requesty in AI Software
See Requesty in the listRequesty is #114 in AI Software by our score.
