AI in the product
LLM API
Call large language models from your product.
Ask your AI about this, with this page as the source:ChatGPT ↗Claude ↗Perplexity ↗
The real choice
Which provider, and whether to go through a gateway. Going direct to a provider gets its newest models and features first and the lowest latency; a gateway gives you one key and one bill for hundreds of models, with fallbacks when a provider is down, for a small markup or a proxy you run. The code that calls the model — an AI SDK or agent framework — is its own decision.
Pick by situation
Tap the ones that are you — the tools that fit light up below.- You want the strongest general models and the widest ecosystem of examples and integrations
OpenAI API
Claude
Gemini API - Token cost dominates your budget, for example high-volume batch or agent workloads
DeepSeek
Groq
Mistral AI - You want to try or mix many models with one key and one bill, and fall back when a provider is down
OpenRouter - You want that routing and spend control but self-hosted, with keys and logs on your servers
LiteLLM
The contenders
Grouped by the side of the choice they answer, not ranked. Open a row for when to use it, the trade-off and what makers say.
ToolBest for
Model provider5
OpenAI APIapiThe broadest model lineup — text, images, speech, embeddings — and the largest ecosystem.
734+481 open source$10GPT-6 Luna34.1M/wk+61% vs npm
Use it whenYou want one account that covers text, images, speech and embeddings, with the most examples to copy from.
Trade-offFrequent model and API changes to keep up with, and no option to run its hosted models elsewhere.
llms.txtIn Lovable, ReplitUsed by 2dto3D, 8base, Agentplace.io, AgentX, Agnost AI and 729 more · in 481 open-source projects.
We use GPT-5.6 to power the Basedash AI data analyst. It's incredibly intelligent and, with the right harness, allows us to rank #1 on BI Bench for solving real-world BI scenarios.
ClaudeapiCoding, long documents and agent-style tool use.
251+285 open source$10Claude Haiku 5.531.6M/wk4.1× vs npm
Use it whenYour feature is an agent that calls tools, writes code or works through long documents.
Trade-offNo image generation or embedding models of its own, so those need a second provider.
llms.txtIn Lovable, ReplitUsed by 8base, Accordio AI, Agentcard, Agentplace.io, AgentX and 246 more · in 285 open-source projects.
Taskade Genesis works beautifully with Claude. It brings strong reasoning and long-range thinking to any Genesis app, helping systems plan, adapt, and make sense of complex inputs.
Gemini APIapi · free tierVery long context and multimodal input, with a free tier to start.
99+227 open source$28Gemini 3.1 Flash-Lite19.2M/wk6.5× vs npm
Use it whenYou feed in whole documents, video or audio, or want to start without paying.
Trade-offFree-tier prompts may be used to improve Google's products, so paid tier is the one for user data.
In Lovable, ReplitUsed by 1752vc Pitch Deck Analyzer, 8base, Accorata, Agentplace.io, Appear on AI and 94 more · in 227 open-source projects.
Powers all agent conversations on Konfide. Fast, cost-effective, handles unlimited concurrent chats. Every user message goes through Gemini. Chose it for speed and quality at scale.
Mistral AIapi · free tierA European provider whose API covers chat, coding, OCR and speech models.
15+69 open source$6.00Ministral 3 (3B)7.1M/wk6× vs npm
Use it whenYou want an EU-based provider, or document OCR alongside text models.
Trade-offA smaller ecosystem of third-party integrations than the largest providers.
llms.txtIn ReplitUsed by Internet.io, Kuration AI, MeetMinutes, Officely AI, Omnifact and 10 more · in 69 open-source projects.
Talespinner uses Mistral Large as a writing model. It is great for writing novels with more mature topics, since it doesn't censor your writing that much.
DeepSeekapiLow per-token prices on strong reasoning and coding models, with OpenAI- and Anthropic-format endpoints.
11+24 open source$27deepseek-flash1.9M/wk4.9× vs npm
Use it whenToken cost dominates your budget, for example high-volume batch or agent workloads.
Trade-offServed from China, which matters for some customers' data requirements; peak-hour prices are double off-peak.
Used by Genstore.ai, IFTTT, MGX (Now Atoms), Read & Give by Bono, Shipable AI by CNTXT AI and 6 more · in 24 open-source projects.
DeepSeek delivers serious reasoning at a fraction of the cost of comparable models. We chose it because powerful AI shouldn't have a premium price tag attached to every use case.
Open models, hosted1
Groqapi · free tierVery low-latency inference on open models, for real-time features.
39+62 open source$6.75GPT OSS 20B1.8M/wk3.9× vs npm
Use it whenResponse speed matters more than having the most capable model, as in voice or autocomplete.
Trade-offOnly the open models it chooses to host, and no frontier closed models.
Used by Antispace, Automaticall, Bebop.ai, ClarityUX, Extra Thursday and 34 more · in 62 open-source projects.
Shoutout to Groq — the lightning fast inference engine running Voicr's AI under the hood. Without Groq's speed, that under 3 seconds promise wouldn't be possible.
Gateway (many providers, one API)2
OpenRouterapi · free tierTrying many models from many providers with one API key and one bill, without opening an account at each.
61+31 open source—2.1M/wk4.8× vs npm
Use it whenYou want to compare models quickly or offer users a model picker.
Trade-offA fee on top of provider prices, and your traffic and data pass through a third party.
llms.txtIn ReplitUsed by Actx0, AI Emaily, Breadcrumb, ChikitAI, Clado and 56 more · in 31 open-source projects.
Talespinner uses OpenRouter to allow for switching between different language models. They make it super easy to do this, because they offer one API format for all language models.
LiteLLMlibrary · open source · free tierRunning your own OpenAI-compatible proxy in front of your own provider keys, with retries, routing and per-key budgets.
21+97 open source—148.1M/wk
Use it whenYou want keys and request data to stay on your infrastructure.
Trade-offYou deploy, monitor and upgrade it yourself; SSO and audit logs are in the paid enterprise tier.
llms.txtMCPUsed by CamelAI, Crossnode, Fume, PandaProbe, PandaProbe Cloud and 16 more · in 97 open-source projects.
We use LiteLLM mainly so we’re not tied to a single model provider. Makes it super easy to switch models or add fallbacks when running AI services for clients.
Cost: the cheapest plan that fits input tokens 50 million tokens, output tokens 10 million tokens, from list prices. Try your own numbers →
Cost as you grow
Each contender's cheapest usable plan as usage rises.
Who switches to what
Public pull requests on GitHub since Oct 2024 whose title says "X to Y" — real code changes, by developers in general rather than makers only. Pick a flow to see its pull requests.
OpenAI API → Gemini API179+ PRs · Compare →- feat(dh): switch listup web search from OpenAI to Geminigrowthhackerssnu/admin · 2026-10-01
- Switch production thumbnails from OpenAI to Geminikevin4relocate/TripCaution · 2026-09-30
- feat: move OpenAI and Google defaults to GPT-6 and Gemini 3tag1consulting/ai-pr-review · 2026-09-29
- Migrate from OpenAI to Google Gemini for embeddings and LLManiilhr/agentic-ai-rag-chatbot · 2026-09-29
- fix: preserve image parts in OpenAI to Gemini responsesmoeacgx/NewAPI-Mao · 2026-09-27
- Updating from OpenAI to Gemini in llm.py following the potential toke…harpand87/banking-support-agent · 2026-09-27
- Switch the assistant LLM from OpenAI to Geministariik/Mia · 2026-09-26
- port(upstream): #4274 - disambiguate colliding tool call ids in openai to geminibloodf/durindoor · 2026-09-25
- feat(gemini): convert OpenAI Embeddings to Gemini embeddingstbphp/gpt-load · 2026-09-25
- Switch AI provider from OpenAI to Google GeminiOricrew/service-scheduler · 2026-09-24
Gemini API → Groq152 PRs · Compare →- College MVP: switch LLM provider from Gemini to Groq (sole provider)akshat2685/pathmind · 2026-10-01
- Switch idea rating/roadmap generation from Gemini to GroqLaKhWaN/startup-game · 2026-09-29
- Switch in-app AI provider from Gemini to Groqjubayerjuhan/cognivo · 2026-09-24
- Migrate from Gemini API to Groq APIHushnudbek-s-organisation/ScholarBridgeAi · 2026-09-23
- feat(ai): switch AI diagnostic engine from Google Gemini to GroqSHOEBILL04/Garij · 2026-09-21
- feat(gateway): switch the non-technical summary from gemini to groqMicroTodoSuite/microservice-app-slack-approval-gateway · 2026-09-21
- Use Gemini as Jarvis's reply engine, fall back to Groq on failuredeepseatrader18/deepsea-dashboard · 2026-09-19
- changed the model provider from google gemini to groqRit2002/placeintel · 2026-09-19
- Switch AI Assistant from Gemini to Groqevery1hatestaha-png/busniessOS · 2026-09-16
- Switch formatting/ front & back matter conversion from Gemini to Groqlndat18/production-legal-qa-rag · 2026-09-16
OpenAI API → Claude127 PRs · Compare →- Migrate AI provider from OpenAI to Anthropic Claudesawthunaing/stn-job-hunting-dashboard-with-ai · 2026-10-01
- Preserve readable context when switching owned OpenAI to ClaudeJJLiebig/codex · 2026-10-01
- Add OpenAI Codex plugin marketplace to Claude Code project settingsbrandohermoso47-rgb/cliente-vault-web-app · 2026-10-01
- Switch Wardrobe from OpenAI to Claudenjarrar/wardrobe-najeeb · 2026-09-28
- Switch receipt extraction from OpenAI to Claudebelindahbl/receipt-telegram-bot · 2026-09-26
- Switch AI copy from OpenAI to Claude Haiku 4.5fareeha-s/shouldve-walked · 2026-09-26
- Phase 1: Switch AI provider from OpenAI to Anthropic Claudeanfaznil/tabletalk-ai · 2026-09-23
- fix(thinking): route OpenAI and level-based reasoning models to claude-adaptive on Claude endpointshkbulbul/polyrouter · 2026-09-23
- Switch FAQ replies from OpenAI to Anthropic Claudeimranaisolutions/Agentic-Ai · 2026-09-23
- Switch AI provider from OpenAI to Claude, and hand redirects to AJR Core (4.1.0)andrew-ajrwebdesign/ai-seo-assistant · 2026-09-22
Gemini API → Claude108+ PRs · Compare →- fix(refine): session-end refine uses gemini first, falls back to claude on availability failuresJohnGavin/llm · 2026-10-01
- Add Gemini as an alternative AI research provider to ClaudeNikros07/Eye-of-Horus · 2026-10-01
- Bring Codex CLI, Gemini CLI and Grok Build up to Claude Code's featuresoddessentials/agent-guild · 2026-09-30
- docs: migrate governance reviews from Gemini to ClaudeKaancodm/GeniusNew · 2026-09-27
- Add Gemini as a free alternative to ClaudeHarshith017/maxxtempo · 2026-09-26
- Add Gemini MCP server exposing ask_gemini tool to Claudekeshav321-kg/fintech · 2026-09-25
- Switch summary generation from Gemini back to Claudevarunsagar20/module2-DB · 2026-09-23
- Switch `update-github-info` from unpriced Gemini pin to priced Copilot Claude modelnaveedausaf/skills-agentic-workflows-that-read-the-room · 2026-09-17
- Switch AI content generation from Gemini to the Anthropic Claude APIkniepertsebastian-spec/learning · 2026-09-17
- chore(ci): triage every new issue with Gemini, and hand approved ones to Claude with claude:gohectorcanaimero/orch · 2026-09-14
Claude → Gemini API105+ PRs · Compare →- Switch the chatbot from Claude to Gemini (free tier)anelisa-lab/AI-Budgeting · 2026-09-29
- Migrate AI agent from Claude to Google GeminiVaishnavi6267/social-media-engagement-agent · 2026-09-28
- Switch model provider from Claude to Google Geminidhruhisheth/rightsline · 2026-09-26
- Switch AI backend from Anthropic Claude to Google Geminimasasukexx-tech/en-marketing-lp · 2026-09-22
- feat(agentcc): route Claude Agent SDK requests to Vertex Geminifuture-agi/future-agi · 2026-09-21
- Switch the LLM goal compiler from Claude to Geminiabimundayat26/twin-bank · 2026-09-20
- Switch the AI engine from Anthropic Claude to Google Gemini (free tier)itzvxie/demo-site · 2026-09-20
- Member 1: Switch AI agent from Claude to Gemini (free tier)Dibyajyoti882/MetricMind · 2026-09-20
- Migrate voice assistant from Claude to Google Gemini APIfeyyazonkal33/iddaa-analiz-app · 2026-09-17
- Switch AI rule parsing from Claude to Geminiananyaprabhakarm/AI-Backtesting-Agent · 2026-09-14
Gemini API → OpenAI API90+ PRs · Compare →- gemini fallback to openai fixesmaximhq/bifrost · 2026-10-02
- Switch LLM backend from Gemini to OpenAI gpt-4o-miniPiSquareLabs/PathShala · 2026-09-30
- perf(gemini): Migrate Gemini to OpenAI formattisfeng/Easydict · 2026-09-29
- fixes foreign-function-call-id issue for fallback from gemini to openaimaximhq/bifrost · 2026-09-29
- Switch the model calls from Gemini to OpenAI gpt-5.4-miniMishaMatcharashvili/trip-io · 2026-09-25
- refactor: migrate AI insights service from Gemini API to OpenAI APITkYaLa34/AI-Investment-Research-Platform- · 2026-09-23
- Switch AI provider from Gemini to OpenAI, add dialogue lessons, fix e…aaryashah1010/Fluentify · 2026-09-22
- migrated from gemini to openai llmmohitgarg196/CodeVitals-AI · 2026-09-21
- Fix Gemini adapter to support OpenAI SDK client formatarvindg4u/APIKeyRotator · 2026-09-20
- Fix Gemini chat model + image fallback to OpenAIfabelanorma3010/How--to-wite--a-fiction-book · 2026-09-05
Claude → OpenAI API89 PRs · Compare →- fix: round-trip Claude tool_use to OpenAI tool_callsks1686/peaproxy · 2026-09-27
- Switch receipt AI scan from Claude to OpenAIianselmi/cardmaster · 2026-09-26
- Migrate from Claude to OpenAI API for synthesis toolsarkhash0309/research-paper-intelligence-pipeline · 2026-09-25
- Switch Alora Chat from Claude to OpenAIvaniabhagania/Alora · 2026-09-24
- Migrate all Claude resume tailoring to OpenAIrfahamedansari/job-application-assistant · 2026-09-23
- feat(claude): map output_config.format to OpenAI response_formatrouter-for-me/CLIProxyAPI · 2026-09-23
- feat(claude-boxes): allow codex egress to the OpenAI auth and API hostsLanternfyre/home-lab · 2026-09-22
- [synthmind-bot] Security Researchers Use Anthropic Claude to Hack OpenAI — 2026-09-18administrakt0r/AIautonomousBlogMY · 2026-09-18
- Switch LLM provider from Anthropic Claude to OpenAI GPT-4o.sji-soumik/ai-assessment · 2026-09-11
- fix: clamp Claude Code egress to OpenAI and Gemini (RUN-1515)NeuralTrust/TrustGate · 2026-09-10
Groq → Gemini API64 PRs · Compare →- Fall back from Groq to Gemini text when the model is retired or rate-limitedcyangjr/marketplace-scout · 2026-10-01
- Switch Lumi chat from Groq to Google Gemini 2.0 FlashThabisoCollinSengane/Pulsify · 2026-09-20
- Switch deepeval CI judge from Groq to Gemini 3.1 Flash-Litetruongpx396/agent-core-demo · 2026-09-17
- fix: route Site Agent plain Groq calls to Gemini fallbackakamanim/khasroy · 2026-09-13
- Move the AI features from Groq to Geminibogdan0089/fastapi-ecommerce-backend · 2026-09-11
- Switch AI grading/recommendation backend from Groq to Geminicrabb-beltran/de-dojo · 2026-08-31
- feat(songs): switch screenshot extraction from Groq to Geminialesmo30/song-shift · 2026-08-24
- switched groq api call to gemini apiNafisaTasnimR/Keepify · 2026-08-23
- Migrate LLM backend from Groq to Geminimuizzusman/youtube-content-engine · 2026-08-21
- Switch LLM and transcription from Groq to Geminiabhijeetmishra2104/sonicscribe-app1 · 2026-08-18
Before you choose
How to approach it
Put model calls behind one function in your code so switching providers is a one-line change. Log prompts, responses and cost from day one. Set spending limits on every provider account.
Common mistakes
- Shipping with no spending limit or per-user rate limit, so one abusive user or a retry loop runs up the bill overnight.
- Hard-coding one model name across the codebase, so moving to a newer or cheaper model means a search-and-replace and a round of surprises.
- Putting the provider API key in client-side code, where anyone can copy it from the browser.
Other options
Real choices most makers here won't need to weigh.
Azure OpenAIThrough your cloudOpenAI models inside an Azure account, with regional and EU/US data-zone deployments.23 makers' products
Together AIOpen models, hostedPer-token access to a wide catalog of open-weight models, with fine-tuning and dedicated GPUs from the same vendor.12 makers' products
Grok APIModel providerGrok models with built-in web and X search, plus image, video and voice APIs.6 makers' productsMiniMaxModel providerLow per-token prices on open-weight coding and agent models, with speech, image, video and music models in the same account.6 makers' products
Fireworks AIOpen models, hostedServerless and dedicated inference on popular open-weight models, with fine-tuning and batch jobs at half the serverless price.5 makers' products
Vertex AIThrough your cloudGemini and partner models such as Claude on Google Cloud, with its regions, IAM and billing.3 makers' products
PortkeyGateway (many providers, one API)A gateway that also handles guardrails, prompt management and request logs in one dashboard.2 makers' productsEden AIGateway (many providers, one API)One API key and bill for LLMs and non-chat AI — OCR, speech, translation, vision — across many providers, with an EU endpoint.2 makers' productsDeepInfraOpen models, hostedLow per-token prices on a large catalog of open-weight models, plus embeddings, speech and image models, behind an OpenAI-compatible API.2 makers' products
Amazon BedrockThrough your cloudClaude, Llama, Mistral, Amazon Nova and other models inside an AWS account, billed with the rest of your AWS usage.1 makers' product
ReplicateModel providerOpen-source image, video and audio models without running GPUs yourself.
Vercel AI GatewayGateway (many providers, one API)One key and one bill for hundreds of models at provider list prices, with fallbacks and budgets, built into the AI SDK.
BifrostGateway (many providers, one API)A self-hosted, open-source gateway written in Go that routes one OpenAI-compatible API to many providers with failover, load balancing and caching.Qwen APIModel providerQwen models plus DeepSeek, Kimi and GLM under one Alibaba Cloud account, with OpenAI- and Anthropic-format endpoints.Kimi APIModel providerKimi K3, an open-weight model with 1M-token context, and coding models behind OpenAI- and Anthropic-format endpoints.GLM APIModel providerGLM coding and agent models through Z.ai, with free Flash models and Anthropic-format access for Claude Code.Doubao APIModel providerByteDance's Seed models alongside its Seedream image and Seedance video models, with an OpenAI-compatible API.ERNIE APIModel providerBaidu's ERNIE models through Qianfan, with an OpenAI-compatible endpoint.Hunyuan APIModel providerTencent's HY models plus DeepSeek, GLM, Kimi and MiniMax on one TokenHub key, in OpenAI and Anthropic formats.Cerebras Inference APIOpen models, hostedVery fast output — thousands of tokens per second — on a small set of open-weight models.Meta Model APIModel providerMeta's Muse Spark models for agents and coding, with a 1M-token context, multimodal input and OpenAI- and Anthropic-format endpoints.Xiaomi MiMo APIModel providerVery low per-token prices on open-weight models with a 1M-token context and image, audio and video input, through OpenAI- and Anthropic-format endpoints.Decided alongside
What the 919 makers' products here chose for their other decisions.
Frontend Framework314 products
Next.js260
React35
Astro21
Next.js260React35Astro21
Resend
SendGrid
Postmark
Amazon Web Services
Cloudflare
Google Cloud PlatformDatabase245 products
Supabase117
Firebase34
MongoDB32
Supabase117Firebase34MongoDB32Payments217 products
Stripe177
Lemon Squeezy16
Paddle12
Stripe177Lemon Squeezy16Paddle12Product Analytics101 products
PostHog76
Mixpanel20
Amplitude9
PostHog76Mixpanel20Amplitude9
Plausible AnalyticsAuthentication78 products
Clerk59Supabase6
Auth04
Clerk59Supabase6Auth04