BaseAI — The Web AI Framework
The easiest way to build serverless autonomous AI agents with memory. Start building local-first, agentic pipes, tools, and memory. Deploy serverless with one command.
Evidence: README emphasizes server-side only API keys and warns against exposure; but no specific least-privilege implementation or audit. User confirmation mechanism not documented. Data flow transparency: README describes pipe run and stream events, but not data storage, transmission, or sharing. Sensitive data handling: API key handling tips, but no other sensitive data. Dependency security: dependencies listed, but no security audit or vulnerability scan evidence. External effects: framework calls external LLM APIs, but no side effects or user consent. Rollback: not mentioned. Source attribution: authors listed, but not verified. Deductions: lack of implementation details and evidence.
Evidence: README and package.json are consistent, describing framework purpose and installation steps. Dependency availability: dependency list exists, but version availability not verified. Failure messages: README example has error event listener, but no detailed error handling docs. Deductions: dependency availability and failure messages lack concrete evidence.
Evidence: README targets developers, provides multiple use cases (local dev, deployment). Capability boundaries: describes pipes, tools, memory, but no explicit limits. Trigger precision: pipe run triggered by user message, but no detailed trigger conditions. Environment fit: supports Node.js >=18, but no other environments. Deductions: capability boundaries and trigger precision insufficient.
Evidence: README provides clear directory structure and installation steps. Install notes: npx baseai init etc. Naming stability: no version naming strategy. Examples and FAQ: code examples, but no FAQ. Known limitations: not mentioned. License: Apache-2.0 clear. Versioning changelog: not mentioned. Maintenance responsibility: contributing guide mentioned, but maintainers not explicit. Deductions: missing versioning changelog and known limitations.
Evidence: Output usability: streaming output example, clear output format. Marginal value: framework offers memory and tool integration, unique value. Cost benefit: no performance or cost data. Deductions: cost benefit lacks evidence.
Evidence: Claims in README (e.g., 'easiest way') not supported. Cross-source: no other sources to verify. Fact-inference separation: README mixes facts and marketing language. Deductions: lack of verifiable evidence.
- Repository is archived; consider Langbase AI Primitives as alternative.
- No security audit or vulnerability scan evidence; dependency security unknown.
- API key management requires caution; ensure server-side only.
What does this agent do, and when should you use it?
BaseAI is a TypeScript-first framework for building serverless and composable AI agents with memory and tools. It enables local development of AI agent pipes, integrating agentic tools and memory (RAG), then deployment to serverless with a single command. The repository provides CLI commands for initialization, pipe creation, and running a dev server, plus the @baseai/core library for programmatic use. The project is now archived in favor of Langbase AI Primitives, which offer more flexibility with APIs and SDKs.
BaseAI CLI commands: npx baseai@latest init creates a baseai directory with baseai.config.ts, memory, pipes, and tools subfolders. npx baseai@latest pipe interactively scaffolds a pipe configuration in TypeScript, allowing specification of model, parameters, system prompt, tools, and memory. npx baseai@latest dev starts a local server. The @baseai/core package exports Pipe and getRunner for integration. The Pipe class takes a config object and its run() method accepts messages, returning a stream that can be consumed via event listeners for content, errors, and end events.
- A developer wants to quickly prototype an AI agent with custom system prompts and tool calling, using the CLI to scaffold the project.
- A team building a RAG-powered agent can leverage BaseAI's memory feature for local indexing and retrieval.
- An organization aiming to deploy AI agents as serverless APIs benefits from the one-command deployment to Langbase.
- Node.js developers seeking streaming responses in their apps use the Pipe class and getRunner to handle async events.
- Developers evaluating framework vs. primitives can learn from BaseAI's implementation when considering Langbase.
What are this agent's strengths and limitations?
- TypeScript-first development offers type safety and a good developer experience
- Local-first development with integrated memory and tools accelerates iteration
- One-command serverless deployment simplifies operations
- Supports multiple LLM providers including OpenAI, Anthropic, Cohere, and others
- Archived project is no longer actively maintained, risking lack of updates
- Dependency on Langbase platform for deployment and API key management creates lock-in
- Configuration of multiple API keys can become complex when using many providers
- Framework constraints might not suit highly customized or dynamically evolving agent requirements
How do you install or deploy this agent?
Requires Node.js. Run npx baseai@latest init in your project directory. Then configure a .env file with LANGBASE_API_KEY and API keys for the LLM providers you intend to use (e.g., OPENAI_API_KEY, ANTHROPIC_API_KEY).
How do you use this agent?
- Initialize with
npx baseai@latest init. 2. Add API keys to.env. 3. Create a pipe withnpx baseai@latest pipeand edit the generated configuration. 4. In your app, import the pipe config, instantiatenew Pipe(pipeConfig), and callpipe.run({ messages })to get a stream. 5. Start the local server withnpx baseai@latest devand run your script withnpx tsx index.ts.
FAQ
Does BaseAI require a Langbase API key?
LANGBASE_API_KEY for local execution and deployment.Can I run BaseAI locally without deploying?
npx baseai@latest dev starts a local server, but you need to configure LLM API keys in .env.