Pricing

Tokens are the universal currency on Free.ai. Every tool costs tokens - chat messages, image generations, voice synthesis, etc. One token roughly equals 4 characters of text. Self-hosted models are cheapest, external models (GPT-4, Claude) cost more tokens per use.

No! Purchased tokens and token packs never expire. Daily free tokens reset every 24 hours.

Every tool deducts tokens. Confirmed accounts get 30,000 free tokens per day that reset every 24 hours. Spend that pool however you like: music, images, chat, voice, transcription and OCR are all uncapped within it. Video is the one exception, one clip every two days on a free account, because a single clip costs us more GPU time than everything else combined. A clip is 30,000 tokens, so it is your whole daily pool on the day you make it. A $5 wallet (200,000 tokens) or the $5/mo Basic plan (250K/mo) removes the video cap, and your daily free pool stays active on top.

Self-hosted chat: actual tokens used (~100/message). Image: 1,000 tokens. Video: 30,000 per clip on most models (14,000 on Wan I2V, 47,000 on HunyuanVideo), then 2,500 tokens per second beyond 12 seconds. 3D: 45,500 per mesh on TRELLIS, 11,000 on TripoSR. TTS: ~1 token per 4 chars. STT: ~4 tokens/second. External models (GPT-6, Claude): varies by model, shown in the model selector.

External models (GPT-6, Claude, Gemini, etc.) are billed at cost × a markup that covers card processing, infrastructure, and platform overhead. The markup drops as you commit to larger token purchases - so the bigger your usage, the closer to raw provider cost your effective rate gets:

Trailing 30-day spend	Markup on external models
Under $500	50%
$500 - $1,999	30%
$2,000 and up	15%

Self-hosted models (Qwen, DeepSeek, Mistral, Llama, SDXL, Kokoro, Whisper, etc.) are flat-rate per call regardless of tier - the tier only affects cost-pass-through on external providers. Your current tier is visible on your account page, and updates automatically at the start of each billing cycle.

Yes! Refer friends and both of you get 30,000 bonus tokens. You can also contribute voice data, training data, or report bugs to earn tokens.

Yes. Cancel anytime from your account page. Your remaining tokens stay in your account. We also offer a 30-day money-back guarantee on all plans.

Yes. Every paid plan includes API access - 10,000 requests a month on Basic and Starter, 100,000 on Pro, 1,000,000 on Business. All 477 tools are available programmatically. See free.ai/api/ for documentation.

Visa, Mastercard, American Express, Apple Pay, Google Pay, and PayPal. Card and wallet payments are handled securely by Stripe; we never see your card details.

Yes. The Business plan ($49/mo) includes team workspaces with a shared token balance. On every plan you can add extra seats, either when you subscribe or later from your account. Each seat comes with its own allowance of tokens, private to the person in it, and your own plan allowance becomes the shared team balance that anyone can draw on once theirs runs out. Adding two or more seats at once discounts every seat in that purchase.

Free models (Qwen 2.5, Kokoro, Whisper, FLUX) run on our own GPU servers - they are fast, private, and cost fewer tokens. Premium models (GPT-4, Claude, Gemini) are hosted by their respective providers and cost more tokens.

Yes. Students and teachers with a verified .edu email get 500,000 free tokens per month plus API keys at the Pro request tier, for every tool that runs on our self-hosted models. Apply at free.ai/edu/.

Contact us at hello@free.ai for enterprise pricing. We offer custom plans with higher limits, dedicated GPU access, and SLA guarantees.