Build, launch, and experiment with applications powered by Jev. Open infrastructure for programmable intelligence.
https://t.co/FTshn8v5Mu
0x3739fe5159c1cf463b2fe696d5a5413905a9d5e1
We just made a major update to ensure launching is easy and smooth for users.
Now, If the agent has a minimum holding, the gateway reads the wallet's balance of the coin before issuing the key and again, cached for a minute, on every call. The launcher can change the minimum on chain at any time.
Market capitalization is neither our priority nor a concern to us.
Jev is already outperforming nearly every free AI offering and ultimately/potentially represents the future of specialized AI products.
We will therefore continue to build. If the community has suggestions, we are more than happy to consider and work on them
Minor update for jevlab
Directly fund the budget with creator rewards via pons
Sweep what the coin has earned on pons into its budget. Anyone can; it costs gas and nothing else.
https://t.co/peRj7yb17R
Jev Labs update.
We’re adding agent templates.
Instead of starting from scratch, builders will be able to choose a base model for things like risk scoring, verification, market analysis, or custom evaluations, then configure it and launch.
We’re bringing the same launch flow you use on the web, configure an agent, attach a token, and go live with a signed endpoint onto mobile.
Coming soon
Jev Labs update.
We’re adding live agent analytics.
Every launched agent will have a clear view of calls made, fees generated, budget remaining, and how much intelligence the token economy has funded.
Recent upgrade for the agents -
You can now chain Jev calls instead of stuffing everything into one question set.
Each stage is still one Jev request: typed questions, probabilities, confidence. After stage A answers, your code (or a simple branch rule) builds the next state and hits stage B. Fees from the agent’s coin still pay every hop.
Response includes the full trace + token usage, so you can see cost against the fee sink
Same launch flow. Same signed config. Now the agent can actually sequence decisions instead of answering once and stopping.
https://t.co/1G5LpgVZWy
We are currently making systematic upgrades to all of our agents
Re-writing the docs to make it as easily understandable as possible
And we are also still yet having problems with finalizing our domain. Bare with us please.
Build, launch, and experiment with applications powered by Jev. Open infrastructure for programmable intelligence.
https://t.co/Q5Oa167vfv
0x3739fe5159c1cf463b2fe696d5a5413905a9d5e1
Did you know agents can be changed post launch?
It signs a new config and the endpoint switches to it, with the hash on the agent's page. The budget, no. Nobody can move a budget except the operator, and only to settle calls.
You configure what it evaluates -->
It gets an endpoint the moment it launches -->
It launches with a coin -->
Every trade pays into the budget -->
The budget pays for jev, then it loops.
Our Jev agent doesn’t write, reason through chains, count, or access outside data you provide.
Use it for quick decisions in software classification, routing, scoring, detection, ranking, verification, and safeguards.
Keep an LLM for generating the actual text.