From belief to useful work
A purpose on the surface.
Clear rules underneath.
Telligence helps people fund recurring AI inference for a project. A revnet coordinates the support. A dedicated vault holds the compute backing. The creator puts an API key to work.
01
Support
People contribute to the project’s revnet.
02
Activate
Its compute allocation is converted into backing.
03
Think
Verified backing provides a daily inference allowance.
04
Keep going
The allowance renews each day while the backing and service are available.
What your contribution does
The revnet accepts funding on Base, with VVV as its accounting and reserve asset. Normal Revnet issuance and market buybacks stay in place. Supporters receive participation in the revnet; a fixed production split goes to the project's policy wrapper.
The wrapper can cash out its production allocation for VVV and send the proceeds to the compute vault. Cash-out taxes, available reserves, market routing, and execution affect how much VVV arrives. A 40% production split is a share of tokens produced, not a promise that 40% of money contributed becomes compute.
The vault stakes VVV, locks the resulting sVVV for DIEM, and stakes that DIEM. It owns every part of this position. The website shows provider credit only after it has been observed and verified.
Supporting a purpose does not enforce how its creator uses inference, and it does not automatically share the creator's future revenue. Project revenue reaches the revnet only if someone pays it in.
What “humming” means
A project is humming when its backing has been activated, the service is enabled, and a recent provider observation shows spendable daily credit. Its key may remain useful without new contributions. Fundraising activity and token prices are not a measure of available inference.
Venice credit renews at 00:00 UTC. Unused daily credit does not roll over. Different models consume it at different rates. The API can pause when the day's allowance is exhausted, during a provider outage, or while backing is being recovered.
The dashboard shows verified capacity and today's remaining credit. Missing observations stay unknown.
A useful key, a separate vault
The creator signs in to create, limit, and revoke project API keys. Applications use a compatible inference endpoint. A hosted authentication signer connects to Venice as the vault; it has no authority to withdraw the vault's backing.
Key limits and verified provider credit constrain API usage, along with any existing project cap. Compromise of the upstream signer can still consume the project's provider-visible credit until it is effectively revoked. Compute access and asset authority are separate, but both need care.
Understand the way out
Compute backing is not part of the liquid reserve supporters can immediately cash out. Unwinding a position follows the provider's sequential cooldowns and the vault's recovery policy. Cooldowns can change; recovery is not instant.
Recovered VVV returns to the revnet through its balance-return path. That benefits current holders according to the revnet's rules, rather than promising a refund to the original contributors. Every project links to its revnet, wrapper, and vault for inspection.
The compute service depends on Venice, including its accounts, model availability, credit accounting, and upgradeable staking infrastructure. Fixed Revnet terms do not remove those dependencies.