Sovereign is free. Hosted is fast.
FallForge itself is free MIT. What you pay for (if anything) is convenience: hosted CORS-proxying, done-for-you deployment, or human concierge review of the output.
Local · Sovereign · £0
Runs entirely in your browser. BYOK to any of six LLM providers (or use free WebLLM Botler). Zero data hits our servers.
- Full FallForge 5-phase pipeline
- All 5 router blends available
- ZIP download
- Optional GitHub push (BYOK token)
- Free forever · MIT · fork it
Hosted · £29 one-off
Same pipeline, faster. Cloudflare Worker handles CORS-blocked scraping and deeper crawls. Automatic GitHub Pages deploy.
- Everything in Local
- CORS-free fetching for any site
- Deeper crawl (up to 50 signal pages)
- Automatic Pages enable
- One-off charge · nothing recurring
Pro concierge · £199 one-off
Human reviews the FallForge output before delivery. Tunes the kernel. Calibrates autopilot. Runs the sim on your behalf. You get a ready-to-run instance.
- Everything in Hosted
- Human kernel review + polish
- Autopilot pre-calibrated
- 30-min onboarding call
- 7 days follow-up support
Post-install LLM budget.
These are the recurring costs of the LLM you point FallHub at (paid direct to the provider, never to us). Pick the blend that fits.
| Blend | Providers | £/mo |
|---|---|---|
| Full Frontier | Anthropic Claude Opus 4.8 across the board | £80–£300 |
| 80/20 (default) | Claude for reasoning · Groq Llama for bulk | £20–£60 |
| 50/50 | Balanced across frontier + open | £8–£25 |
| Full Open | Groq Llama 3.3 70B across the board | £4–£12 |
| Local only | WebLLM Llama 3.1 8B (in browser) | £0 |
Rough monthly ranges assume a small SMB with 200-800 messages/month. You can change blend anytime from within FallHub.
White-label FallForge for your clients.
Agencies and consultants: deploy FallForge on your own domain, keep the £199 concierge margin, use your own branding. £999/month via the FallHub Agency tier includes unlimited client installs.