TokenFans Desktop

Turn idle model capacity into reliable supply.

TokenFans Desktop helps you bind a device, set daily limits, and start sharing after you approve the supply. Status, heartbeat, and revenue stay visible in the local console.

TokenFans console dashboard
TokenFans app icon
Early access Contributor desktop beta
Daily limit 500-1000M / day
Launch mode One-click sharing
Runtime Agent + Bridge Core
Settlement USD / MTok

Download

Download the desktop build for your platform.

TokenFans Desktop is available for macOS, Windows, and Linux. Choose the build for your device, then request an invite to join the contributor network.

macOS Apple Silicon Best for always-on Mac mini, Mac Studio, and Apple Silicon laptops. Download DMG
macOS Intel For older Mac devices with Intel processors. Download DMG
Windows x64 For contributor workstations and long-running machines. Download EXE
Linux x64 For always-on devices and server-style desktop environments. Download AppImage

Contributor flow

Start sharing in three steps, without learning the network layer.

01

Bind identity

Enter an invite code, email, and device name. Existing bindings can be restored with the local token.

02

Choose supply

Select Claude, OpenAI, Gemini, regional models, local proxy mode, or hosted API key mode.

03

Start sharing

The desktop app starts Agent and Bridge Core, then reports status, heartbeat, and supply authorization.

Console language

The website uses the same language as the desktop console.

The website keeps only the contributor concepts that matter: bind identity, choose supply, set daily limits, start sharing, and view revenue.

Console
The desktop home screen shows sharing status and key metrics immediately after launch.
Share model capacity
Plainly describes what contributors provide, without forcing technical component names.
M tokens / day
Daily limits are shown in millions of tokens to avoid misleading tiny-scale labels.
TokenFans supply settings screen

Network operations

Operations handles review, scheduling, and settlement.

TokenFans manages contributors, devices, supply grants, revenue, and payouts. Platform services handle gateway routing, protocol forwarding, metering, and failover.

Provider catalog Supply catalog

Organizes capacity by model provider, account type, supply mode, and daily limit.

Supply grant Contributor grant

After authorization, operations reviews whether supply enters the schedulable pool.

Projection Supply projection

Approved supply is projected to base accounts and tunnel nodes, with repair and rollback paths.

Revenue Revenue ledger

Rates are configured by provider, model, token type, account type, and supply mode.

Security

Review first, schedule second, disable supply when needed.

TokenFans production paths are built around private network isolation, Base URL allowlisting, supply grants, heartbeat monitoring, alert snapshots, and emergency stop.

Tailnet Private network with ACLs and grants
Credential Local proxy first; hosted keys require explicit approval
Health Offline state, heartbeat issues, and bridge failures are tracked
Rollback Remove supply from scheduling and revoke devices when required

Early access

Invites are open for early contributors.

If you have a stable model subscription, CLI environment, or shareable API key capacity, apply to join the TokenFans contributor network.

Invite and contact email: info@tokfans.ai

Request invite

FAQ

Frequently asked questions

Do I need to start Agent manually after sharing begins?

No. The desktop app saves configuration and starts Agent plus Bridge Core when you click start sharing.

Will the platform keep scheduling after I turn sharing off?

No. When a contributor turns off sharing or revokes authorization, the supply becomes unschedulable.

Will API keys be uploaded directly to the platform?

Local proxy mode is recommended by default, so API keys stay on the desktop. Hosted keys require explicit contributor approval and enter encryption, masking, revocation, and audit flows.

Why are daily limits shown in M tokens?

Supply comes from different subscriptions and model sources. M tokens better matches how contributors and operations understand practical capacity.