H2H
Our working platform for AI education
H2H is the desktop app we run in workshops and production engagements: centralized access to OpenAI, Anthropic, and Gemini models, cost control, templates, and full usage visibility - in one interface.
Currently preparing
The latest release is still being fetched from GitHub. Try again in a few minutes, or contact us if you need access right now.
Contact usCapabilities
What it does
Three AI providers in one interface
OpenAI, Anthropic, and Gemini models in the same app. Compare outputs side by side without switching tools.
Zero API keys for end users
All traffic is proxied through your backend. Employees never see keys; keys never end up in Slack threads or screenshots.
Budgeting in internal tokens
The PD token system abstracts per-provider pricing. Admins set company-wide and per-user budgets with no surprises on the invoice.
Prompt template library
Tested prompts for common tasks (guest email, contract review, meeting summary). Drag and drop into a chat.
Real-time admin monitoring
Admins see usage per user, per model, per day. Live team visibility in a workshop, cost control in production.
Signed auto-updates
Every release is signed with a minisign key. The app verifies the signature before installing - no manual downloads after the first.
System requirements
- Windows 10 or later (64-bit)
- Internet connection for AI model access
- License key from a workshop or your admin
Frequently asked
- Not currently - H2H is paired with workshop licenses. If you want your team to run it independently, get in touch and we'll set up a dedicated instance.
- We currently ship only a Windows build because that's what our clients use. A Mac build goes on the roadmap when there's demand.
- Conversations are stored in your backend (PostgreSQL, EU-hosted). Nothing persists on OpenAI or Anthropic servers beyond the API call itself.
- Only you and your organization's admins. Every query is checked against companyId via JWT - multi-tenant isolation by default.