Multi-agent chats, zero-latency experiments, infinite ideas.
Welcome to ChatSparty, your playground for exploring multi-agent AI conversations.
We’re building a space where multiple LLM-powered agents can chat, debate, plan, and build ideas together in real time. Whether you’re testing workflows, simulating team decision-making, or just watching your agents argue about who should make coffee, ChatSparty helps you explore what’s possible with local, private, fast multi-agent experimentation.
- Multi-Agent Systems: Create and run multiple AI personas with unique roles and goals.
- Local-First: Use Ollama and local LLMs for privacy and speed.
- Workflow Testing: Simulate and refine agent workflows with persistent memory layers like Mem0.
- Open Source: Hackable, extendable, and community-friendly.
We believe multi-agent systems are the future of AI. Teams of small, specialized AIs working together can create more dynamic, trustworthy, and effective systems than single, monolithic models.
ChatSparty exists to:
✅ Test agent-to-agent communication patterns.
✅ Explore emergent behaviors in teamwork and planning.
✅ Let developers prototype multi-agent workflows quickly.
Interested in exploring multi-agent workflows, building new personas, or integrating with tools like Mem0 for persistent memory? We’d love to collaborate.
- ⭐ Star the repo to follow updates.
- 🛠️ Open issues for feature requests and discussions.
- 🧑💻 Contribute code, prompts, or ideas.
Stay tuned for updates, experiments, and new agent configurations.
ChatSparty Team