Any provider, your keys.

Pluely speaks to any OpenAI-compatible endpoint — 23 providers documented below, from frontier labs to your own laptop. Keys are encrypted on your device, usage is unmetered, and the free plan covers all of it.

Download free Read the setup guide

23 providersKeys encrypted on-deviceUnmetered — free plan covers it

OpenAI

Connect your OpenAI API key and run GPT models directly in Pluely's overlay — your billing, your rate limits, no middleman. Every screenshot question, live meeting answer, and document summary goes straight from your machine to OpenAI and back.

Connect OpenAI

Anthropic

Connect your Anthropic API key and run Claude models in Pluely's overlay — Sonnet for everyday speed, Opus-class models for the hard questions. Claude's long context and careful writing style make it the model many people trust for work that gets read by other humans.

Connect Anthropic

Google AI Studio

Grab a free API key from Google AI Studio and run Gemini models in Pluely — no credit card, no billing setup, and a free tier generous enough to cover real daily use. The fastest zero-cost path from download to answers.

Connect Google AI Studio

OpenRouter

One OpenRouter key unlocks essentially every model on the market — GPT, Claude, Gemini, Llama, DeepSeek and hundreds more — all usable from Pluely's overlay. If you like trying new models the week they drop, this is the key to have.

Connect OpenRouter

Groq

Groq's custom LPU hardware streams hundreds of tokens per second — connected to Pluely, full answers appear almost before you finish asking. For live meetings, where every second of silence counts, raw speed is a feature.

Connect Groq

Mistral AI

Mistral's La Plateforme offers strong models with a real free tier and EU data residency — the natural choice if your work or your clients need European data handling, and a genuinely good one even if they don't.

Connect Mistral AI

DeepSeek

DeepSeek's V3 and R1 models deliver near-frontier quality at a fraction of the usual token price — the key for people who ask their overlay a hundred questions a day and don't want to think about the bill.

Connect DeepSeek

xAI

Connect your xAI key and run Grok models in Pluely — fast reasoning with unusually fresh knowledge, over every app you use. Grok's personality also makes it the most fun model to have hovering over your screen.

Connect xAI

Together AI

Together hosts the open-model universe — Llama, Qwen, DeepSeek, Mixtral — behind one OpenAI-compatible API with per-second serverless billing. The practical way to run big open models without owning a GPU.

Connect Together AI

Fireworks AI

Fireworks serves open models with enterprise speed, compliance, and reliability — the choice when Pluely is part of how you work, not an experiment, and the endpoint simply has to be up.

Connect Fireworks AI

Perplexity

Perplexity's Sonar models answer from the live web with citations — connect your key and Pluely's overlay becomes a search engine that reads your screen. Ask about the thing in front of you and get an answer backed by sources, not vibes.

Connect Perplexity

Azure OpenAI

If your organization runs OpenAI models through Azure, point Pluely at your deployment and keep every request inside your company's compliance boundary — same GPT models, corporate rules satisfied.

Connect Azure OpenAI

Cohere

Cohere's Command family is tuned for business tasks — grounded answers, summarization, tool use — and its Compatibility API plugs directly into Pluely. A workhorse choice for document-heavy days.

Connect Cohere

Ollama (local server)

Beyond the CLI integration, Ollama also serves an OpenAI-compatible API on localhost — add it as a custom provider and every model you've pulled appears in Pluely. Answers generated on your own silicon, nothing leaving the machine.

Connect Ollama (local server)

LM Studio

LM Studio runs open models locally with a friendly desktop UI and serves them on an OpenAI-compatible port — the least technical path to a fully offline Pluely. Browse models, click download, connect, done.

Connect LM Studio

Cerebras

Cerebras inference runs on dinner-plate-sized chips and it shows — thousands of tokens per second on frontier open models. Connected to Pluely, even long answers land in a blink. Free tier available.

Connect Cerebras

DeepInfra

DeepInfra serves the big open models at some of the lowest per-token prices anywhere, with per-request billing and no minimums — the budget key for making Pluely your all-day habit.

Connect DeepInfra

Novita AI

Novita hosts popular open models behind an OpenAI-compatible API with startup-friendly pricing and login credits — an easy, cheap key to get Pluely answering on DeepSeek or Llama today.

Connect Novita AI

Moonshot AI (Kimi)

Moonshot's Kimi K2 family tops open-model agentic benchmarks — a trillion-parameter MoE that reasons, plans, and writes like a frontier lab's flagship. Connect a Moonshot key and run Kimi inside Pluely's overlay.

Connect Moonshot AI (Kimi)

Alibaba Cloud (Qwen API)

Alibaba's Model Studio serves the full Qwen lineup — Max, Plus, Coder, VL — over an OpenAI-compatible endpoint, with free quotas for new accounts. First-party Qwen, in Pluely's picker.

Connect Alibaba Cloud (Qwen API)

Z.ai (GLM)

Z.ai's GLM models pair huge context windows with strong reasoning and agentic skills — and their coding-plan pricing is famously aggressive. Connect a key and run GLM in Pluely's overlay.

Connect Z.ai (GLM)

NVIDIA NIM

NVIDIA's build.nvidia.com serves optimized open models over standard APIs with generous free credits for developers — another zero-cost key that drops straight into Pluely.

Connect NVIDIA NIM

Hyperbolic

Hyperbolic serves frontier open models at aggressive prices on marketplace GPU capacity — BF16 precision, low prices, OpenAI-compatible, Pluely-ready.

Connect Hyperbolic

Your keys, one overlay.

Install Pluely, paste a key once, and every provider you connect is a tap away in the model picker — over meetings, documents, and everything on your screen.

Download free Prefer a CLI?

Free plan forever · Invisible on screen shares · No bot joins your calls