codelit.io

What It Is

Codelit.io positions itself as an Architecture-First SDLC and Agent Workflows platform, emphasizing an end-to-end process from specification to production: designing systems, simulating failures, generating code, and deploying. The clearest capabilities found in the crawled content fall into two categories: first, an Agent Workflow Builder for designing production-grade AI Agent workflows before wiring up implementations; second, API v1, which can generate system architecture JSON from text and analyze GitHub repositories.

Core Capabilities and Ecosystem

Its Agent Workflow Builder covers elements such as Skills, MCP servers, specialist agents, tool permissions, model routing, eval harnesses, guardrails, and deployment targets. This suggests it is not merely a diagramming tool, but an attempt to bring Agent runtime behavior, permissions, evaluation, and deployment into a unified specification. On the API side, /api/generate accepts a 3–500 character system description and returns structured architecture results including title, description, nodes, edges, and metadata. /api/github/analyze can analyze owner, repo, and branch, returning detected files, technology patterns, directory structure, and source files for architecture generation. The documentation provides cURL, JavaScript, and Python examples, so the barrier to developer integration appears relatively low.

Pricing and Deployment

The crawled text only shows an entry point to the Pricing page and a View Pricing link, without specific plans, prices, free quotas, or enterprise plan details. The API uses an OpenRouter key, and the documentation states that rate limits depend on the upstream provider account. It supports Authorization Bearer or an apiKey in the request body. No information was found about open source vs. closed source status, self-hosting, data retention, security, or compliance, so enterprises should conduct additional due diligence before adoption.

Pros and Cons

The main strengths are its focused concept and the structured JSON output from the architecture generation API, making it suitable for embedding into internal developer platforms, documentation systems, or architecture review workflows. GitHub analysis can also help infer technology patterns from existing repositories. The drawbacks are that public information is limited, and pricing and deployment models are not transparent. Models and rate limits depend on OpenRouter, so upstream stability and costs need to be assessed independently. Supported languages and frameworks can only be inferred from examples, with no complete compatibility statement available.

Who It’s For and Access from China

It is suitable for software architects, platform engineering teams, AI Agent developers, and teams looking to automate architecture diagram generation and repository analysis. Access from China cannot be determined from the available text, and OpenRouter-related services may have uncertainties around domestic network access and payments. It is recommended to test domain availability, API latency, and payment routes in practice. If you need more mature alternatives for diagrams or architecture documentation, consider Mermaid, PlantUML, Structurizr, Eraser, Miro, or Lucidchart.