Portkey
AI gateway with routing, caching, guardrails and observability across model providers.
why this verdict
Keep paying — OpenAI has not replaced it
OpenAI's version overlaps, but does not finish the dev tools job, so this one still earns its $49/mo.
- A named launch, not a vibe named, unlinked
- OpenRouter and cloud AI gateways offering routing and fallbacks with less setup. OpenAI, August 5, 2025 — no announcement link recorded yet.
- How much of the job it covers not the job editorial call
- Parts of it. The job still needs the tool to get finished.
- Is there a free way to do it? yes
- 3 of 3 listed replacements have a usable free tier: OpenRouter, LiteLLM and Helicone.
- What the call is worth $588/yr
- $49/mo at the entry paid tier — $588 a year per seat.
- Threatened by
- OpenAI
- Since
- August 5, 2025
- List price
- $49/mo
- Per year
- $588
The backstory
Portkey sits between an application and every model vendor it calls, handling retries, fallbacks, semantic caching, budget limits and guardrails from one control plane. The category is crowded and OpenRouter took the simple routing use case. Portkey's differentiation is governance: enterprises with dozens of teams calling models need per-team budgets, prompt versioning and an audit trail, which is a procurement requirement rather than a developer preference, and it is the harder thing to replicate.
Escape hatches
Simplest way to reach many models through one API
openrouter.ai open_in_newOpen-source proxy with a compatible interface
litellm.ai open_in_newLighter option focused on logging and cost
helicone.ai open_in_new3 of 3 replacements have a usable free tier.
