Open source · WASM sandboxed · Manifest governed
Permissioned autonomy.Modular intelligence.
A secure, plugin-based execution framework for autonomous agents on Web3. Designed for composability, governed by manifest.
using System;
using System.Threading.Tasks;
public class JuliaAgent
{
private readonly IPluginManager _pluginManager;
private readonly IManifestValidator _manifestValidator;
public JuliaAgent(IPluginManager pluginManager, IManifestValidator manifestValidator)
{
_pluginManager = pluginManager;
_manifestValidator = manifestValidator;
}
public async Task<bool> ExecutePlugin(string pluginId, object parameters)
{
try
{
var plugin = await _pluginManager.LoadPlugin(pluginId);
if (!_manifestValidator.ValidateManifest(plugin.Manifest))
{
throw new InvalidOperationException("Invalid plugin manifest");
}
return await plugin.Execute(parameters);
}
catch (Exception ex)
{
Console.WriteLine($"Error executing plugin: {ex.Message}");
return false;
}
}
}Features
Everything an agent needs, nothing it doesn’t get to touch.
Open source
Free forever. Run it anywhere.
Read the manifest spec, scaffold a plugin, and ship your first sandboxed agent in an afternoon.
Read the docsPlugin system
Capabilities an agent can be granted, one plugin at a time.
Julia plugins extend agent capabilities — from sending transactions and querying chain state, to validating zero-knowledge proofs and solving strategic games through embedded game theory engines.
public class GameTheorySolver
{
private int[,] payoffMatrixA;
private int[,] payoffMatrixB;
public GameTheorySolver(int[,] a, int[,] b)
{
payoffMatrixA = a;
payoffMatrixB = b;
}
public List<(int, int)> FindPureStrategyNashEquilibria()
{
var equilibria = new List<(int, int)>();
int rows = payoffMatrixA.GetLength(0);
int cols = payoffMatrixA.GetLength(1);
for (int i = 0; i < rows; i++) // Strategy A
{
for (int j = 0; j < cols; j++) // Strategy B
{
bool isBestForA = true;
bool isBestForB = true;
// Check if player A has a better response in row
for (int k = 0; k < rows; k++)
{
if (payoffMatrixA[k, j] > payoffMatrixA[i, j])
{
isBestForA = false;
break;
}
}
// Check if player B has a better response in column
for (int l = 0; l < cols; l++)
{
if (payoffMatrixB[i, l] > payoffMatrixB[i, j])
{
isBestForB = false;
break;
}
}
if (isBestForA && isBestForB)
{
equilibria.Add((i, j));
}
}
}
return equilibria;
}
}Nash equilibrium
This plugin searches for pure strategy Nash equilibria in a two-player normal-form game. Each player picks a strategy, and the solver checks whether that choice is mutually optimal.
- 01
For every cell (i, j) in the payoff matrices
Walk the full strategy space. Brute force is exact here, and small games are the common case.
- 02
Player A has no better response than i, given B plays j
Scan A's column. If any row pays more, this cell is not stable for A.
- 03
Player B has no better response than j, given A plays i
Scan B's row. If any column pays more, this cell is not stable for B.
- 04
Both hold — (i, j) is a Nash equilibrium
Neither player gains by deviating alone. Record the pair and continue.
Best suited to small strategy spaces. For mixed strategies or large games, reach for numerical or linear programming methods — plugin documentation.
Pricing
Scale your agent operations.
Production features and extended control for teams, protocols and DAOs.
Advanced runtime
- Multi-agent orchestration
- Custom plugin registry
- Enforced manifest policies
- Agent metrics + observability
- Web dashboard (beta)
- Dedicated Discord support
Free
Open source forever
No fees. No lock-in. Run anywhere you want.