Free Credits & Grants
AI Credits Directory
Curated list of free compute credits, startup grants, and API incentives from every major AI and cloud platform.
Not sure which credits you qualify for?
Answer 3 questions and get a personalised match — takes 90 seconds.
Amazon Web Services
AWS Activate
AWS Activate is Amazon's flagship startup credit program, offering up to $100,000 in AWS credits for qualifying startups.
Google Cloud
Google Cloud for Startups
Google Cloud for Startups combines up to $200,000 in GCP credits with access to Google's full AI stack -- Vertex AI, Gemini API, BigQuery, Cloud Run, and more.
Quick check
Early-stage startup (Seed to Series A preferred for standard tier)
Less than $5M in Google Cloud credits previously received
Cursor
Cursor AI Editor
Cursor is the AI-native IDE built on VS Code. The free tier gives every developer 2,000 code completions and 50 premium model requests per month. Cursor Pro ($20/mo or $16/mo annual) unlocks unlimited completions, access to Claude 3.5 Sonnet, GPT-4o, and the Composer multi-file context tool. Many engineering teams now subsidize Cursor for all engineers — it measurably reduces time-to-ship.
Quick check
Free tier available to any developer — no company or funding requirement
Pro plan requires a credit card ($20/mo or $192/yr)
Ramp
Ramp Corporate Card
Ramp is the corporate card and spend management platform built for tech companies. Unlike traditional corporate cards, Ramp underwrites based on your company's cash position, not founders' personal credit scores. New companies earn $500 in cashback rewards. Ramp integrates natively with QuickBooks, Xero, NetSuite, and Sage — making it the default choice for startups managing heavy AWS, GCP, or Azure spend who want automated expense reconciliation.
Quick check
US-registered business entity (C-Corp, S-Corp, or LLC)
Business bank account with at least $25,000 in operating funds
Mistral AI
Mistral AI La Plateforme
Mistral AI's API provides access to Mistral Large, Mistral Small, Codestral, and Pixtral. EU-hosted for GDPR compliance. New accounts receive free credits; competitive pricing makes it cheaper than GPT-4 for most tasks.
Quick check
Valid email — no company required
EU hosting included; GDPR-compliant by default
CoreWeave
CoreWeave
CoreWeave provides specialised GPU cloud infrastructure for large-scale AI workloads. Startup program offers up to $100K in compute credits for companies in YC, a16z, Sequoia, and other qualifying accelerators.
Quick check
Portfolio company of a Tier 1 VC or accelerator (YC, a16z, Sequoia, Lightspeed, etc.)
Pre-Series B startup with active AI workloads
Modal Labs
Modal Labs
Modal Labs provides a $30/month free tier plus startup credits, representing a fundamentally different approach to AI infrastructure -- serverless GPU compute that developers consistently describe as transformative.
Quick check
Free tier: any developer with a GitHub account
Startup credits: portfolio company of YC, a16z, Sequoia, or other partners
Groq
Groq Cloud
Groq Cloud provides the fastest inference currently available -- often 5-10x faster than OpenAI or Anthropic on identical model sizes -- making it the go-to choice for latency-sensitive applications.
Quick check
Valid email address — no credit card required
Developer account; no company required
Emergent Labs
Emergent
Emergent is the AI-native platform for building, deploying, and iterating on production-grade full-stack web apps — entirely through conversation. No local setup, no DevOps configuration, no boilerplate.
Quick check
Any developer, founder, or AI builder can start for free — no credit card required
Free tier: up to 5 apps and 50 AI-assisted builds per month
Deepgram
Deepgram
Deepgram provides the fastest and most accurate speech-to-text API. New accounts receive $200 in free credits — enough to transcribe ~55 hours of audio. Faster-than-real-time transcription with speaker diarization included.
Quick check
Valid email — no credit card required for free credits
$200 is permanent (no expiry), not a trial
Vercel
v0 by Vercel
v0 is Vercel's AI UI generator. Describe a component or screen in plain English and v0 produces production-ready React + Tailwind + shadcn/ui code that you can copy or open directly in StackBlitz. Free tier includes 200 credits/month — roughly 20–30 full UI generations. Premium ($20/mo) unlocks 5,000 credits and priority queue. Engineering teams use v0 for rapid prototyping, design-system bootstrapping, and turning Figma descriptions into working code in seconds.
Quick check
Free Vercel account required (GitHub, Google, or email signup)
No company or funding requirement — open to all developers
Replit
Replit AI
Replit combines a cloud IDE with Replit Agent — an AI that scaffolds, writes, debugs, and deploys full-stack apps from a single prompt. The free tier includes Ghostwriter completions. Replit Core ($25/mo) unlocks the full Replit Agent for autonomous development, custom domains, private Repls, and extra compute. The platform is especially powerful for prototyping, hackathons, and solo founders who want zero DevOps.
Quick check
Free tier available to anyone — no credit card required
Core plan ($25/mo) requires a credit card
Lambda Labs
Lambda Cloud GPU
Lambda provides GPU cloud instances at 50–70% less than AWS or Azure. New accounts receive $10 in free credits. H100 instances from $2.49/hour — the most cost-effective option for large model fine-tuning.
Quick check
Valid email and credit card (required to claim the $10 credit)
Startup program: qualifying early-stage companies receive additional credits
Microsoft Azure
Microsoft for Startups
Microsoft for Startups (Founders Hub) offers up to $150,000 in Azure credits -- and since Azure is the backbone of OpenAI's infrastructure, this is the most direct path to GPT-4o and o1 API credits at scale.
Quick check
For-profit startup, not publicly traded
Not raised more than Series A (<$5M institutional funding for entry tier)
DigitalOcean
DigitalOcean GPU Droplets
DigitalOcean GPU Droplets put NVIDIA H100 and A100 GPU instances in the hands of developers without enterprise negotiation. New accounts receive $200 in free credits valid for 60 days — enough for meaningful fine-tuning experiments or batch inference runs. Per-second billing means you pay only for active compute time. DigitalOcean's simpler control panel and predictable pricing make it a developer favourite vs. AWS or GCP for straightforward GPU tasks.
Quick check
New DigitalOcean account — $200 credit applied automatically
Valid credit card required for identity verification (not charged if within credit)
StackBlitz
Bolt.new
Bolt.new is a browser-native full-stack development environment powered by AI. Built on StackBlitz WebContainers, it runs Node.js entirely in the browser — no local setup, no Docker. Describe any web app and Bolt generates, runs, and iterates on a complete codebase with React, Vue, Astro, or vanilla JS and a backend API in minutes. The free tier includes daily credits for generation. Bolt Pro ($20/mo) unlocks higher daily limits and GitHub export sync.
Quick check
Free tier — no account or credit card required to try it
Account required to save and revisit projects
Anthropic
Anthropic API Credits
Anthropic's Claude Credit Program provides early-stage AI teams with $5,000-$25,000 in API credits, specifically for developers building production applications on Claude 3.5 Sonnet, Haiku, and Opus.
Quick check
Early-stage startup (typically seed to Series A) building a product on Claude
Not already a high-spend Anthropic API customer
LangChain
LangSmith
LangSmith provides LLM tracing, evaluation, and monitoring — essential for debugging RAG pipelines and agent workflows. Free tier includes 5,000 traces/month. Used by teams at Replit, Retool, and Notion.
Fireworks AI
Fireworks AI
Fireworks AI provides fast, production-grade inference for open-source models. New accounts receive $1 in free credits (~1M tokens). Llama 3.1 70B at $0.90/M tokens — among the cheapest for quality inference.
Quick check
Valid email — no credit card required for trial
Pay-as-you-go after trial credits
Helicone
Helicone
Helicone is a drop-in LLM observability proxy with zero-code setup — log every request, cache responses, track costs, and apply guardrails. Free tier handles 100K requests/month. Single line of code change to integrate.
RunPod
RunPod GPU Cloud
RunPod offers on-demand and spot GPU instances starting from $0.19/hour for RTX 4090s. Community Cloud provides the cheapest H100/A100 rates available. New accounts receive $0.50 in free credits to get started.
Quick check
Valid email — minimal verification required
Credit card required after free credits
Weights & Biases
Weights & Biases
W&B is the standard MLOps platform for experiment tracking, model versioning, and dataset management. Free forever for solo developers with unlimited projects. Academic and startup programs offer team features at no cost.
Quick check
Individual plan: any developer, permanently free
Academic: valid institutional email
Stability AI
Stability AI API
Stability AI's API provides access to Stable Diffusion 3.5, SDXL, and Stable Video Diffusion. New accounts receive 25 free credits — enough for 12+ high-quality images or 2 video clips. Industry-leading image generation.
Quick check
Valid email — no credit card required for free credits
Age verification may be required
Baseten
Baseten
Baseten enables one-command deployment of any ML model to production-ready GPU endpoints with automatic scaling. Startup program (via YC and accelerators) provides compute credits. Used for deploying custom Llama, Whisper, and diffusion models.
Quick check
Startup program: portfolio company of YC or partner accelerator
Developer trial available for any user
xAI
xAI Grok API
xAI's API provides access to the Grok-3 family of models with real-time web search capability and a 128K context window. Free tier available with 25 API requests/hour. Grok-3 Mini is competitive with GPT-4o for reasoning tasks at lower cost.
Portkey
Portkey AI Gateway
Portkey is a production LLM gateway with automatic failover, load balancing, semantic caching, and guardrails. Free tier handles 10,000 requests/month. Supports 250+ LLMs via a single OpenAI-compatible API.
Databricks
Databricks Startup Program
Databricks provides a unified analytics platform with Mosaic AI for LLM fine-tuning, RAG development, and ML lifecycle management. The Startup Program offers credits for qualifying pre-Series B companies through VC partner networks.
MongoDB
MongoDB Atlas
MongoDB Atlas is a fully managed cloud database with built-in vector search (Atlas Vector Search). The M0 free cluster includes 512MB storage with no time limit — ideal for AI app metadata and semantic search.
Quick check
Email account — no credit card required for free tier
One free M0 cluster per project
Supabase
Supabase
Supabase provides Postgres + auth + storage + edge functions in one open-source platform. Free tier includes 500MB database with pgvector for semantic search. The go-to backend stack for AI-powered apps.
Railway
Railway
Railway is a deployment platform that eliminates DevOps overhead. Connect a GitHub repo, select a service template — Postgres, Redis, Node, Python, Go — and Railway handles build, deploy, networking, and autoscaling automatically. Every account gets $5 in free monthly credits. The Hobby plan ($5/mo) provides $10/mo in usage credits. Teams use Railway for staging environments, internal tools, and side projects where full Kubernetes orchestration would be overkill.
Quick check
GitHub account required for authentication
Free $5/mo credits apply automatically to all new accounts
Fly.io
Fly.io
Fly.io runs containerised apps in 35+ regions with sub-10ms latency globally. Free tier includes 3 shared VMs and 3GB storage — enough to host a production AI API. GPU instances available for $0.50/hour.
Quick check
Email account and credit card required (card is for identity verification — free tier has no charges)
Free tier is permanent
Neon
Neon Serverless Postgres
Neon provides serverless Postgres that scales to zero between requests. Free tier includes 0.5GB storage and the pgvector extension for semantic search. Branching lets you create database copies instantly for testing.
Runway ML
Runway Gen-3 Alpha
Runway Gen-3 Alpha represents current state-of-the-art in AI video generation. Their credit program offers $15-$125 for developers and creators exploring video AI.
Scale AI
Scale AI (Spellbook)
Scale AI provides data labelling infrastructure and the Spellbook AI platform for evaluating and deploying LLMs. Developer program includes free Spellbook access. Startup credits available for qualifying early-stage companies.
Quick check
Developer program: apply at scale.com/api
Startup program: Series A or earlier with AI-first product
Cerebras Systems
Cerebras Inference
Cerebras runs the world's fastest LLM inference -- using their custom Wafer Scale Engine, they achieve 2,200+ tokens/second on Llama 3.1 70B, making Groq look slow by comparison.
Brex
Brex Business Card
Brex is a business card and financial platform built for startups and high-growth companies. Underwriting is based on your company's cash position — not your personal credit score — meaning pre-revenue founders can get approved on day one. Earn Brex points redeemable for AWS, GCP, and Stripe credits, effectively giving a rebate on cloud bills. Brex also offers built-in expense management, receipt capture, and real-time spend controls for your entire team.
Quick check
US-registered C-Corp, S-Corp, or LLC
Minimum $50,000 in a business bank account (waived for YC/a16z/Sequoia-backed companies)
OpenAI
OpenAI API Platform
OpenAI's API Platform credit programs offer $2,500-$25,000 for developers building on GPT-4o, o1, o3, Whisper, and the broader OpenAI ecosystem.
Quick check
Startup or developer team actively building a product on OpenAI APIs
Typically seed to Series A stage
Together AI
Together AI
Together AI provides serverless inference for 50+ open-source models — Llama 3.1 405B/70B, Mixtral 8x7B, Code Llama, FLUX, and more — at throughput speeds that regularly outperform competing providers on benchmark leaderboards. New accounts get $25 in free API credits. Together AI is the go-to for teams that want open-source model flexibility without the cost and complexity of self-hosting GPU infrastructure.
Quick check
New Together AI account — free to create with GitHub or email
No special eligibility for base $25 credits — available globally
Vultr
Vultr Cloud GPU
Vultr offers NVIDIA A100 and H100 GPU cloud instances across 32 global data centers with straightforward hourly pricing and no long-term commitment. New accounts receive $250 in free credits when signed up through a referral link — enough for several hours of H100 runtime or extended A100 experiments. With sub-minute provisioning, Vultr is a developer favourite for fine-tuning runs, batch inference, and anyone shopping for cost-competitive alternatives to AWS or GCP GPU instances.
Quick check
New Vultr account — $250 credit requires signup via a referral link
Standard signup (no referral) gets $100 in credits
Replicate
Replicate
Replicate makes running AI models as simple as calling an API -- no infrastructure setup, no GPU management. Their credit program gives developers $5-$500 to explore the platform.
Quick check
Anyone can sign up — no specific startup eligibility for the base $5 credits
Startup program ($500) available for early-stage teams — apply at replicate.com/startups
Jasper
Jasper AI
Jasper is an enterprise AI writing platform that goes beyond a generic chatbot. It integrates with your brand voice, marketing campaigns, and SEO workflows to produce consistently on-brand content at scale. The 7-day free trial includes full access to Blog Post Wizard, Social Posts, Campaign templates, and SEO mode — no credit card required. For teams evaluating AI content platforms, Jasper also runs a generous affiliate program (20% recurring commission) making it one of the stronger monetisation opportunities in the AI content space.
Quick check
7-day free trial — open to anyone, no credit card required
Business and Team plans require a payment method
GitHub
GitHub Copilot
GitHub Copilot -- the AI pair programmer that started the AI coding revolution -- offers 30-day free trials plus free access for verified students, educators, and open-source maintainers with popular repositories.
Quick check
FREE for verified students (GitHub Student Developer Pack)
FREE for open-source maintainers with active repositories (1,000+ stars, or invited by GitHub)
Perplexity
Perplexity Pro
Perplexity is an AI search engine with real-time web access that cites every source. Every search produces a synthesised answer with clickable references — making it the go-to tool for developers who need fast, verifiable research. The free tier supports unlimited basic searches. Pro ($20/mo or $200/yr) unlocks Claude 3.5 Sonnet, GPT-4o, unlimited file uploads, image generation, and the Perplexity API for building search-grounded applications.
Quick check
Free tier — works without an account, no signup required
Pro plan: valid email and credit card required
Test Provider
Test Credit Program
A test submission via UI automation
AI21 Labs
AI21 Labs
AI21 Labs offers access to Jamba 1.5 — a hybrid SSM-Transformer model with a 256K context window — and the Jurassic model family. Free $10 trial credit on account creation. Jamba offers competitive performance at lower cost than GPT-4.
Quick check
Valid email — no credit card required for trial
Trial credit available to all new accounts globally
Cloudflare
Cloudflare Workers AI
Cloudflare Workers AI runs AI models at the edge -- on Cloudflare's global network of 300+ data centers -- giving your AI features sub-50ms latency for users worldwide.
Quick check
Any Cloudflare account (free tier sufficient)
Workers subscription required for production scale
Hugging Face
Hugging Face Pro
Hugging Face Pro and compute credits (via ZeroGPU and Hub grants) give access to the world's largest open-source model repository -- 500,000+ models -- plus Spaces hosting and GPU inference credits.
Quick check
Free tier: any developer account
Startup program: early-stage company with VC backing or accelerator membership
ElevenLabs
ElevenLabs
ElevenLabs provides the most natural-sounding AI voice generation. Free tier includes 10,000 characters/month (approximately 10 minutes of audio). Creator and above plans include instant voice cloning from any audio sample.
Quick check
Valid email — no credit card required for free tier
API access available on all paid plans
AssemblyAI
AssemblyAI
AssemblyAI offers accurate transcription plus audio intelligence: speaker detection, sentiment analysis, topic detection, and chapter generation. Free tier includes 5 hours of transcription per month with no credit card.
Pinecone
Pinecone
Pinecone is the leading managed vector database for production AI. The free Starter tier includes one serverless index with up to 2GB of storage — sufficient for most RAG applications and semantic search prototypes.
Quick check
Email account — no credit card required for free tier
One project and one index on free tier
Weaviate
Weaviate Cloud
Weaviate is an open-source vector database with built-in ML models for vectorisation. The free Sandbox tier supports up to 1M objects. Startup program provides 6 months of the Professional tier ($55/month value) for qualifying companies.
Quick check
Sandbox: any developer email
Startup program: early-stage company with demonstrable product
NVIDIA
NVIDIA NIM API
NVIDIA's NIM API provides access to 100+ optimised AI models including Llama 3.1, Mistral, Code Llama, and Stable Diffusion. New users receive 1,000 free inference credits, with models optimised for NVIDIA hardware.
Quick check
NVIDIA developer account (free to create)
Credit card not required for free credits
Cohere
Cohere
Cohere provides NLP APIs for generation, embedding, and reranking — optimised for enterprise search and RAG applications. Free trial tier with generous limits. Startup program provides $1,000 in credits for qualifying early-stage companies.
Quick check
Free trial: any developer account
Startup program: pre-Series A startup with active product