Awesome AI AgentsFinancial & Trading Systems

BlockRunAI/Franklin

⭐ 554 TypeScript repository created 2026-03-22

Franklin Agent is a command-line AI agent from the BlockRun team that carries a cryptocurrency wallet and spends from it autonomously to complete tasks, licensed Apache-2.0 and distributed as a single npm package. The premise is that an agent given a budget and an outcome can decide for itself which models and paid services to call, so the wallet replaces accounts, subscriptions and per-provider API keys: the wallet is the identity. Every paid action is settled through the x402 micropayment protocol against the user's own wallet, priced as provider cost plus a fixed margin, which the project brands as pay-only-for-outcome rather than pay-for-access or pay-per-attempt. Trading is the flagship scenario. The agent buys live market data, drafts trade plans that a human approves before any funds move, pursues goals that span multiple sessions, and keeps a trading journal bound to the wallet. Beyond trading it is used for research and general autonomous tasks. A smart router picks a model per task from more than fifty providers, and the tool runs for free out of the box on a set of open models, with frontier models and paid tools unlocked once the wallet is funded with a small amount of USDC. Setup is a global npm install or a direct npx invocation followed by a setup command for Base or Solana and a balance check; the README documents Node version requirements in detail because older releases crash on the Solana dependencies. The same agent also ships as a VS Code extension with a chat panel, model picker, wallet balance display, image and video generation and inline diff cards, driven by the same funded wallet. Guardrails covered in the documentation include human approval before spending and budget ceilings that stop a run.

https://github.com/BlockRunAI/Franklin

ai-agentautonomous-agentautonomous-economic-agentai-tradingcrypto-nativellm-agentmicropaymentspay-per-actiontypescriptusdcwalletx402cli

Also in Financial & Trading Systems

TauricResearch/TradingAgents

TradingAgents is an open-source multi-agent financial trading framework leveraging large language models to simulate real-world trading firm dynamics and collaboratively make informed trading decisions.

HKUDS/Vibe-Trading

Vibe-Trading is an AI-powered personal trading agent leveraging LLMs for comprehensive algorithmic trading, backtesting, and multi-agent financial market analysis.

Fincept-Corporation/FinceptTerminal

FinceptTerminal is an advanced C++/Qt6/Python-based desktop application providing comprehensive financial analytics, investment research, and real-time trading capabilities with AI agents and exten...

hummingbot/hummingbot

Hummingbot is an open-source framework for building and deploying high-frequency crypto trading bots on various centralized and decentralized exchanges.

ValueCell-ai/valuecell

ValueCell is a community-driven, multi-agent platform that leverages AI to provide comprehensive financial applications including stock research, automated trading, and market analysis across multiple global markets.

Lumiwealth/lumibot

Python library for building, backtesting and live-trading algorithmic strategies, with a built-in LLM agent runtime whose agents query market, filings and macro data tools and can place broker orders.

pskrunner14/trading-bot

A Stock Trading Bot that uses Deep Q-Learning, a form of Deep Reinforcement Learning, to learn and execute trading strategies based on historical stock market data.

whchien/ai-trader

Config-driven backtesting framework for algorithmic trading built on Backtrader, shipping 20+ strategies, multi-market data fetching and an MCP server that lets AI assistants run backtests.