Gateway
Unified API gateway to route, secure, and observe 3000+ LLMs across 1600+ providers.
If you run production AI across many LLMs and need central control, resilience, and cost visibility, Portkey's AI Gateway is a solid pick. Smart routing, caching, and guardrails are concrete wins over DIY proxies. For hobbyists or single-model apps, it's overkill, but for scaling teams, it delivers.
Verified 4d ago · liveness 81/100 · cite: rightaichoice.com/tools/gateway
- Platform teams routing to many LLMs needing central control
- Developers building production-grade AI apps with high availability
- Enterprises needing SSO, RBAC, and compliance (SOC 2, GDPR, HIPAA)
- Teams running AI agents at scale with governance and monitoring
- Hobbyists using a single model with no governance needs
- Teams wanting a fully self-hosted solution without cloud features
- Non-developers seeking a no-code AI builder
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip Portkey's AI Gateway if you only use a single LLM provider, have no need for advanced routing or caching, and prefer a provider-native dashboard without additional cost or complexity.
Going past 10k recorded logs per month on the free tier stops recording, so you lose observability until you upgrade.
Portkey's free tier suits prototyping, while the $49/month Production plan fits small teams needing caching, guardrails, and RBAC. Compared to DIY proxies or competitors like LiteLLM (which is open-source but less feature-rich), Portkey's pricing is moderate; enterprises with compliance needs face custom pricing, which can be higher than alternatives like Helicone.
In short
Gateway — Unified API gateway to route, secure, and observe 3000+ LLMs across 1600+ providers. Best for Platform teams routing to many LLMs needing central control, Developers building production-grade AI apps with high availability, Enterprises needing SSO, RBAC, and compliance (SOC 2, GDPR, HIPAA). Free to start; paid plans from $49/mo.
What's new in Gateway
Checked 4 days agoAcross the latest 3 updates: 3 news mentions.
Why Every Agent Vulnerability is a Trust Boundary Failure
This blog post discusses how agent vulnerabilities arise from trust boundary failures, emphasizing the need for governance.
What MCP Governance Actually Means in Production
Explains MCP governance in production contexts, likely tying to Portkey's agent gateway features.
What's an agent gateway?
Defines the concept of an agent gateway, aligning with Portkey's roadmap.
What people actually say about Gateway — is it worth it?
We ran a structured research pass across product reviews, community discussions, and post-purchase forum threads to surface the patterns vendors won't publish themselves. Below: the recurring strengths, the hidden costs people mention most, and the cohort that consistently regrets adopting this tool.
69 mentions across 5 sources (Reddit, Hacker News, App Store, GitHub, Lemmy) · researched Jul 3, 2026.
- +Unified API for 1600+ LLMs from 250+ providers simplifies multi-model management.
- +Built-in guardrails (deterministic and LLM-based) enhance safety and compliance.
- +Enterprise-grade features: SSO, SOC 2, HIPAA, RBAC, and virtual key management.
- +Generous free tier with open-source core fosters adoption and customization.
- +Smart routing with fallbacks, load balancing, and canary testing improves reliability.
- −Almost no real community feedback—trust relies on vendor claims and GitHub metrics alone.
- −211 open GitHub issues may signal unresolved bugs or slow issue resolution.
- −No direct user testimonials about ease of setup or developer experience.
- −Support quality is unverified—no reviews on response times or helpfulness.
- −Limited visibility into performance under high concurrency or production load.
- • Overage fees if exceeding tier request limits (not explicitly documented in community data).
- • Custom enterprise pricing may require long-term contract or minimum commit.
In users’ own words
“Hey folks, I’m starting two new online businesses and already running into trouble with payment gateways — Stripe banned me multiple times (still not sure why). Here’s what I’m building: 1. A company registration business – Needs to accept one-time payments from clients in India, USA, UK, and Dubai. 2. A dropshipping spy tool (subscription-based) – Needs to handle recurring monthly payments from the same…”
Real posts from independent users, linked to the source — not testimonials we collected.
Viability Score
How well maintained and how widely used is Gateway? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: August 2026
How we score →Key Features
- Unified API for 3000+ LLMs across 1600+ providers
- Conditional routing with configurable rules
- Automatic fallbacks and failover
- Load balancing across models and providers
- Canary testing for new models
- Automatic retries with exponential backoff
- Request timeouts
- Simple and semantic caching with configurable TTL
- Observability: logs, traces, feedback, filters, alerts
- Virtual key management with rotation, revocation, monitoring
- Prompt management with templates, playground, versioning, variables
- LLM and partner guardrails
- Multimodal support: vision, audio, image generation
- OpenAI real-time API recording
- Smart batching via provider batch APIs
About Gateway
Portkey's AI Gateway is an enterprise-grade API layer that unifies access to 3000+ LLMs across 1600+ providers, replacing a tangle of direct integrations with a single control plane. Built for platform and DevOps teams running production AI, it centralizes routing, key management, and observability. You get conditional routing, automatic fallbacks, load balancing, and canary testing to keep apps resilient when models fail or underperform. Automatic retries and request timeouts rescue failed calls, while simple and semantic caching cut latency and cost for repeat requests. The gateway is multimodal by design, supporting vision, audio, and image generation providers, plus recording of OpenAI's real-time API including cost and guardrail violations. It handles advanced workloads like smart batching via provider batch APIs, fine-tuning through the unified API, and file uploads for referencing content in requests. Observability is baked in: logs, traces, feedback, filters, and alerts give end-to-end visibility. Virtual keys stored in Portkey's vault let you rotate, revoke, and monitor usage with granular control. For enterprises, the platform offers role-based access control (RBAC), SSO, and compliance certifications (SOC 2 Type 2, GDPR, HIPAA) on the Enterprise plan. Freemium tiers let you prototype for free, while production starts at a predictable $49/month. The open-source core can be self-hosted, or you can use the hosted service. Portkey positions itself as the control panel for production AI, trusted by Fortune 500s and startups alike. Compared to rolling your own proxy or using provider-native dashboards, Portkey gives you a single pane for multi-model governance and cost control. For simple single-model projects, the overhead may not be worth it, but for scaling GenAI teams, it's a dependable choice.
Behind the Verdict
Portkey's AI Gateway is a strong fit for platform and DevOps teams that juggle multiple LLMs and need centralized control. Its conditional routing and fallback mechanisms directly address the pain of model outages and performance variances. The caching—both simple and semantic—is a concrete money-saver, especially for high-volume workloads with repetitive queries. Virtual keys and budget controls give you granular cost management, and the observability suite (logs, traces, feedback) covers what you need to debug and optimize. However, it's not for everyone. If you're building a hobby project with a single model, the added layer is unnecessary overhead. The free tier's 10k log limit and short retention may frustrate those needing longer visibility without paying. The $49 Production tier is reasonable for small teams, but overage charges for logs can surprise you if you scale quickly. Enterprise features like SSO, VPC, and custom guardrails are locked behind custom pricing, which may be a dealbreaker for some. Compared to alternatives like LiteLLM or Helicone, Portkey offers a more comprehensive feature set with built-in caching, routing, and prompt management, but at a higher cost. For teams that value ease of integration and a single pane for governance, it's a worthwhile investment.
Researching Gateway? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas Gateway actually fits — and what changes day-one when you adopt it.
You need to route requests to the cheapest model that meets quality thresholds, and automatically fallback when a provider has an outage.
Outcome: Configure conditional routing rules and fallbacks in the Portkey dashboard; within a day, your app uses the gateway, improving reliability and cutting costs.
You need to monitor all AI usage across teams, enforce budgets, and ensure compliance with SOC 2.
Outcome: Use virtual keys with budgets and RBAC to control access; the observability suite provides logs, traces, and cost breakdowns, while SSO ensures enterprise compliance.
You want to cache similar user queries to reduce API costs and improve response time for your agent.
Outcome: Enable semantic caching; repeated requests get served from cache, reducing latency and cost, while the gateway still logs all activity for debugging.
Use Cases
- Route AI traffic to cheapest/most reliable model per request
- Automatically failover to backup providers during outages
- Cache semantically similar prompts to cut API costs
- Monitor all LLM calls with logs, traces, and cost breakdowns
- Enforce guardrails to filter harmful content or validate output
- Manage API keys centrally with virtual keys and budgets
- Govern and observe AI agents via MCP/Agent Gateway
Models Under the Hood
as of 2026-08-19
Limitations
- The free tier limits recorded logs to 10k per month with 3-day log retention and 30-day metrics retention; exceeding 10k logs stops recording beyond the limit, though requests are not blocked.
- The Production plan ($49/mo) allows 100k logs with 30-day retention and 90-day metrics, but charges $9 overage per additional 100k logs.
- Advanced security (SSO, data residency, custom guardrails, VPC, private cloud) requires the Enterprise plan.
- The gateway is API/CLI only—no dedicated desktop or mobile app.
as of 2026-08-18
Verification history
We have re-verified Gateway 5 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published Gateway tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Open Source
$0
Ideal for
Self-hosters and developers who want full control over their gateway with no cloud dependency, ideal for prototyping or meeting strict data policies.
What this tier adds
Free, self-hosted, includes universal API, retries, routing, and basic dashboard, but lacks advanced guardrails and semantic caching.
Developer Free Forever
$0/mo
Ideal for
Developers prototyping or testing POCs with up to 10k logs per month, needing basic observability and prompt management.
What this tier adds
Includes universal API, fallbacks, load balancing, simple caching, and 3 prompt templates, with 3-day log retention.
Production
$49/mo
Ideal for
Teams ready to deploy LLM apps in production needing higher log capacity, semantic caching, and RBAC.
What this tier adds
Adds 100k logs/month, unlimited prompt templates, LLM & partner guardrails, semantic caching, and RBAC, plus $9 per additional 100k logs.
Enterprise
Custom
Ideal for
Large organizations with compliance needs, high-volume workloads, and custom security requirements.
What this tier adds
Adds 10M+ logs, custom retention, SSO, private cloud/VPC, SOC2, GDPR, HIPAA compliance, custom guardrails, and dedicated support.
Where the pricing makes sense
The company stage and team size where Gateway's pricing actually pencils out — and where peers do it cheaper.
Portkey's free tier suits prototyping, while the $49/month Production plan fits small teams needing caching, guardrails, and RBAC. Compared to DIY proxies or competitors like LiteLLM (which is open-source but less feature-rich), Portkey's pricing is moderate; enterprises with compliance needs face custom pricing, which can be higher than alternatives like Helicone.
Setup time & first value
How long it actually takes to get something useful out of Gateway — broken out by persona, not the marketing-page minute.
For a developer familiar with APIs, you can integrate Portkey's gateway in about 2 minutes using the provided SDKs (Python, Node.js) by changing the base URL or using the Portkey client. For a full production setup with routing, caching, and guardrails, expect a few hours to a day for configuration. For enterprise onboarding with SSO and custom deployment, it may take a week or more.
Switching to or from Gateway
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Direct OpenAI API: Change base URL to Portkey's gateway endpoint and add your Portkey API key to gain routing, caching, and observability without rewriting code.
- ↗To DIY Proxy: You can migrate by replacing Portkey's endpoint with your own reverse proxy, but you'll lose built-in features like caching and guardrails unless you reimplement them.
Integrations
Resources & Guides
Tutorials & Learning
Tools that pair well with Gateway
Common stack mates teams adopt alongside Gateway, with the specific reason each pairing earns its keep.
OpenRouter Agents
One unified AI API for 200T+ monthly tokens across 500+ models, pay-per-token.
OrcaRouter
Zero-markup AI gateway that grades every prompt and routes it to the best model for cost, quality, or speed.
Crazyrouter
Most cost-effective AI gateway: 300+ models, one OpenAI-compatible key, smart cost-optimized routing
Featured Head-to-Head Comparisons
Gateway vs Temporal Ai
Choose Temporal if you need reliability and statefulness for AI agents or multi-step workflows. Choose Gateway if you manage many LLM providers and need routing, guardrails, and cost optimization. They solve different problems—pick based on whether you need durable execution or unified LLM access.
Gateway vs Audioeye
Gateway and AudioEye serve entirely different needs. Gateway is for teams scaling LLM-based apps, needing centralized routing, guardrails, and observability. AudioEye is for organizations requiring web accessibility compliance. Choose Gateway if you manage AI agents/models; choose AudioEye if you need ADA/WCAG compliance.
Gateway vs Push Security
Don't compare apples to oranges: Push Security is a browser security platform for stopping AiTM phishing and AI data loss, while Gateway is an LLM API gateway for routing, guardrailing, and monitoring AI usage. If you're a security team worried about browser-based attacks, choose Push Security. If you're a platform team managing multiple LLMs and need cost/performance optimization with guardrails, choose Gateway.
Alternatives to Gateway
View allOpenRouter Agents
One unified AI API for 200T+ monthly tokens across 500+ models, pay-per-token.
OrcaRouter
Zero-markup AI gateway that grades every prompt and routes it to the best model for cost, quality, or speed.
Crazyrouter
Most cost-effective AI gateway: 300+ models, one OpenAI-compatible key, smart cost-optimized routing
Frequently Asked Questions
Used Gateway? Help shape our editorial sentiment research.