webui · phone app · proxy — for your Hermes agent
run your agent from your phone or laptop — your way
agent-foundations gives you the webui, the phone app, and the proxy that routes your chats. Pick how you want to host the rest — bring your own box, let us run it, or host in your own Cloudflare. Same UX, three shapes.
Bring your own
Tier 1
$5/ month
We give you the webui, phone app, and proxy. You run the rest — on Hetzner, AWS, your laptop, anywhere.
- →BYO LLM key (OpenAI, Anthropic, OpenRouter, anything OpenAI-compatible)
- →BYO Honcho workspace (subscribe or self-host; we never see the data)
- →BYO runtime (your host, your uptime, your network policy)
- →Webui PWA + phone install + branded dashboard included
We host
Tier 2
$15/ month + usage
Same UX as Tier 1, but we run the Hermes runtime (SmolVM microVM) + Honcho VM + nightly R2 backups for you. Pay for what you use.
- →Everything in Tier 1
- →Hermes runtime runs as a SmolVM microVM on our ESXi (Firecracker-backed, tenant-isolated, $0.005/microVM-hr)
- →Dedicated Honcho VM on our ESXi (tenant-isolated, $0.50/GB-mo memory)
- →Nightly R2 snapshots of your Honcho DB ($0.01/GB-mo, 7-day retention)
- →Live public URL (e.g. honcho.agent-foundations.com) via Cloudflare Tunnel
- →Restore from any snapshot within 24h — SLA-backed
You host on Cloudflare
Tier 3
$5/ month
Same UX as Tier 1, but the runtime lives in your own Cloudflare account. You pay CF directly.
- →Everything in Tier 1
- →Hermes runs as a CF Container or Worker in your Cloudflare account
- →You pay Cloudflare directly (we never see the bills)
- →Best for: developers who already live in CF, want zero infra on our side
All tiers include the webui PWA, phone app, proxy, dashboard, brand skin, and support. You pay your LLM tokens directly to your provider — never marked up by us.
What every tier includes
- →The webui — installable as a PWA, branded for agent-foundations, runs on your phone and laptop
- →The phone app — the same webui, added to your home screen, push-notified when new chat arrives
- →The proxy — routes chat requests from the webui to your runtime, streams responses back, handles auth
- →The dashboard — token usage, request counts, error rates, runtime status, billing
- →The brand — your own skin if you want, our skin by default
What we never do
- ×Markup LLM tokens — billed by your provider, never by us
- ×See your data — prompts, replies, memory entries, runtime logs stay in your tier
- ×Lock you in — every tier is reversible; downgrade or upgrade by changing one env var