Best for
- Developers looking to build scalable AI agents in TypeScript
- Applications needing persistent memory and context isolation
- Users wanting to integrate MCP servers into agent workflows
D
Daydreams
Community
A TypeScript AI agent framework for building commerce agents with composable contexts and MCP integration.
GitHub source repository ↗
25FMRS · D
Daydreams is an innovative agent framework offering composable contexts and MCP integration, but it is deprecated with focus shifting to commercial applications.
Daydreams is a TypeScript agent framework for building commerce agents. It introduces a composable context architecture, allowing isolated workspaces to be combined for complex behaviors. The framework supports persistent memory, native MCP integration, type-safe actions, and an extension system. Daydreams also offers Dreams Router, a universal AI gateway with x402 pay-per-use USDC micropayments. Note: The repository has deprecated its agent framework part; core focus is now on Agentic Commerce, with recommendations to use lucid-agents and the Pi agent framework.
Tools
The tool list has not been reviewed yet.
Setup
Install core packages via npm: npm install @daydreamsai/core @ai-sdk/openai zod. Then follow the documentation to create agents.
Fit and risk
Not for
- Developers seeking an actively maintained agent framework (this repo is deprecated)
- Users not interested in agentic commerce features or Dreams Router
- Projects requiring official support or stable APIs
Required permissions
- Requires access to configured AI model providers (e.g., OpenAI)
- May need file system access (via MCP)
- May require network access if using specific extensions
Risks and side effects
- Project is deprecated; features may be outdated and unmaintained
- Reliance on third-party services (AI models, MCP servers)
- Misconfiguration can lead to runtime errors
Troubleshooting
- Ensure dependencies are correctly installed: @daydreamsai/core, @ai-sdk/openai, zod
- Check AI model API keys for validity
- Refer to documentation and Discord community for support
Use cases
Build e-commerce agents for customer service, order processing, and inventory management
Create AI assistants with multi-context support
Integrate external tools and services via MCP
Supported clients
Supported clients have not been confirmed yet.