LiteLLM
Open-source proxy that exposes a hundred model providers behind one OpenAI-style API.
why this verdict
Keep it — OpenAI has not replaced it
OpenAI's version overlaps, but does not finish the dev tools job, so this one is still worth keeping open.
- A named launch, not a vibe named, unlinked
- Providers converging on an OpenAI-compatible API, which is the problem LiteLLM exists to solve. 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
- 2 of 3 listed replacements have a usable free tier: OpenRouter and Vercel AI SDK.
- What the call is worth nothing to cancel
- No paid entry tier tracked, so there is no subscription to cancel.
- Threatened by
- OpenAI
- Since
- August 5, 2025
- List price
- free
- Per year
- —
The backstory
LiteLLM normalises every model provider into a single interface, so swapping vendors is a string change rather than a refactor, and its proxy adds keys, budgets and logging for teams. The existential joke is that as providers converge on OpenAI-compatible endpoints, the translation layer matters less. In practice the proxy is what people actually deploy: rate limits, spend caps and per-team keys are the parts that do not converge, and self-hosting keeps prompts inside the perimeter.
Escape hatches
Hosted equivalent with no infrastructure to run
openrouter.ai open_in_newMore governance and guardrails for larger orgs
portkey.ai open_in_newProvider abstraction at the framework level
ai-sdk.dev open_in_new2 of 3 replacements have a usable free tier.
