The most generous free tier in the industry. No credit card required.
All models are available on the free tier. Each model's rate is a multiplier that follows what the model costs us: free (0x) for Llama 3.3 70B, Nemotron 3 Super, Gemma 4 31B and MythoMax 13B; under 0.01x for the smallest paid models; around 0.08x for DeepSeek V4 Flash; around 0.4x-0.5x for Grok 4.3, Grok 4.5 and Claude Sonnet 5.5; about 1.1x for Claude Opus 5.5; and up to about 5.75x for Claude Fable 5. The in-app model picker always shows the live multiplier before you send anything, and the full list, model by model, is at api.rfab.ai/api/tokens/rates.
We aim for about a 50% margin overall once the free tier and the free daily tokens are paid for, well below the industry norm of 80% to 95%.
Story Pass (optional): $13/month makes Grok 4.3 free in stories and companions, up to a daily allowance. Agents, images, video and voice stay pay-as-you-go.
All image models available on free tier. Costs shown in platform tokens per image.
| Image Model | Quality | Speed | Cost per Image |
|---|---|---|---|
| Flux Schnell | Good | Ultra-fast | 300 tokens |
| Anime Art Diffusion | Good | Fast | 2000 tokens |
| Flux Dev | High | Medium | 2500 tokens |
| SDXL | High | Medium | 2500 tokens |
| Realistic Vision V5 | High | Medium | 2500 tokens |
| Flux Pro | Premium | Medium | 4000 tokens |
| Dreamshaper | Premium | Medium | 4000 tokens |
| NSFW Models | Premium | Medium | 4000-5000 tokens |
| Ideogram V3 | Elite | Slow | 9000 tokens |
| Service | Cost | Notes |
|---|---|---|
| Phone Calls (Vapi) | 10,000 tokens/min (~$0.20/min) | Real-time voice via OpenAI Realtime API |
| Email (AWS SES) | Minimal | Two-way email integration |
| SMS (Twilio) | Standard Twilio rates | Text messaging |
| WhatsApp (Twilio) | Standard Twilio rates | WhatsApp messaging |
| Web Search (Tavily) | Included in tokens | Internet search capability |
Most AI companion platforms (Candy AI, OurDream AI, Character.AI Plus, Replika Pro) sell flat-rate "unlimited" subscriptions. Flat-rate unlimited is only affordable when the chat runs on a small, cheap, unnamed proprietary model — so "unlimited" quietly means "unlimited access to the cheapest AI the operator could host," while everything expensive is metered in opaque coins. On several major platforms, a failed generation still consumes your coins.
Reality Fabricator's pricing is designed to be the maximally honest structure available in this industry:
BYOK ("Bring Your Own Key" — we call it "Pay Us Nothing") lets you add your own OpenAI, Anthropic, Google, xAI, or OpenRouter API keys under Settings → Your API Keys. Turns on your keys are billed by the provider directly at raw cost and Reality Fabricator charges nothing. Your keys are never pooled, never cycled, and used only for your own requests. No other AI companion platform offers this.
No. Billing fails closed: if a generation fails and the provider doesn't bill us, you aren't billed. Coin platforms that charge for failures are the reason we put this in writing.
Because "unlimited" in this industry is a pricing trick — it only pencils out on a small hidden model, with everything else coin-metered. We show real per-model prices instead, and give you free 0x models, BYOK, and local Ollama execution for genuinely unlimited paths that don't require lying about what the AI costs.
Yes! You get 150,000 tokens on signup and 25,000 tokens daily, forever. No credit card required. No hidden fees. Free 0x models (Llama 3.3 70B, Nemotron 3 Super, Gemma 4 31B, MythoMax 13B) are completely free with unlimited usage.
You can purchase more tokens via Stripe, or wait for your daily 25,000 token reward. You can also keep using the free 0x models with unlimited usage, or add your own API keys (BYOK) and pay us nothing.
50,000 tokens = $1. Exact package prices are on the tokens page; what a turn costs depends on the model (shown in the picker) and our published margins (3.8x model cost on text-to-text, 2x on agents and voice).
Yes! Every model — including GPT-5.6, GPT-5.4, Claude Fable 5, Gemini 3.1 Pro, and Grok 4.5 (the successors of GPT-4, GPT-4o, and Claude 3.5 Sonnet) — is available using your free daily tokens. 25,000 daily tokens goes a long way on the budget and standard tiers.
No subscription is required. The free tier is forever free and pay-per-use is optional. The one subscription we offer is optional: the $13/month Story Pass, which makes Grok 4.3 free in stories and companions up to a daily allowance.
ChatGPT Plus costs $240/year for a handful of models. Reality Fabricator is free with 15+ models under their real names, autonomous agents, phone calls, and email. Save $240/year.
Phone calls cost 10,000 tokens per minute (about $0.20/min), billed by actual usage — shown up front like everything else.
You get 25,000 tokens daily, automatically. This resets every 24 hours. No limit on how many days you can claim.
150,000 tokens instantly + 25,000 daily forever
Save $119-$359/year vs competitors
Get Started FreeNo credit card required • All features included • No subscription required