Workflow Builder
https://github.com/vercel-labs/workflow-builder-template- Category
- AI Agents
- Rank
- No. 1357Tools index
Previous survey · No. 1362 ·
- Pricing
- Open Source
- Type
- TOOL
- Builder
- vercel-labs
- GitHub
- 1.2k stars
- Date
About
Visual AI workflow automation template — drag-and-drop builder for chaining LLM agents.
What it does
Workflow Builder is a deployable starter for creating, running, and monitoring automated workflows. Users can assemble flows visually or generate them from natural-language prompts, connect supported services, persist definitions in PostgreSQL, inspect execution logs, and turn completed flows into executable TypeScript.
Why it's ranked here
This is a strong foundation for teams that want ownership of both the editor and execution stack. It combines authentication, persistence, code generation, run history, webhooks, scheduled triggers, and numerous service actions. The result is broader than a canvas demo, though adopting it still means operating an application.
What's good
The visual model does not trap workflows inside the editor. It produces type-safe TypeScript with real integration calls, logging, and error handling. Credentials are fetched at runtime from stored integration references, keeping secrets out of logged step parameters. Anonymous work can also migrate when a user links an account.
Tradeoffs
Setup requires Node.js, pnpm, PostgreSQL, authentication configuration, and an AI Gateway key for prompt generation. Several integrations add their own credentials and operational dependencies. The workflow engine dependency is a beta release. This is a private application template, so teams must deploy, configure, and maintain it themselves.
How to use it well
Use it when a TypeScript team needs a self-hosted starting point for internal automation, customer-facing workflow products, or flows that must become inspectable source code. It fits projects that value editable integrations and execution records. It does not replace a fully managed automation service that removes deployment, database, and credential administration.
Technical notes+
README.md documents the React Flow canvas, Workflow DevKit execution, REST endpoints, PostgreSQL persistence, Better Auth, and generated TypeScript using the "use workflow" directive. package.json defines a private Next.js 16 and React 19 application, pnpm scripts, Playwright end-to-end commands, plugin discovery, and workflow 4.0.1-beta.17. lib/api-client.ts processes streamed JSONL workflow mutations and exposes clients for workflows, integrations, users, execution, and webhooks. lib/workflow-store.ts implements Jotai state, debounced autosave, history, selection, and trigger-node deletion protection. lib/workflow-codegen.ts traverses graph connections and converts node references into generated expressions. lib/credential-fetcher.ts resolves credentials server-side by integration ID, while lib/workflow-logging.ts writes step and workflow completion data through Drizzle.
Observed
- License
- Apache 2.0
- Primary language
- TypeScript
- Packaging
- Private pnpm-managed Next.js application template
- Interfaces
- Browser-based visual editor and REST API endpoints
- Deployment support
- One-click Vercel deployment or local Node.js 18+ setup with PostgreSQL
- Testing surface
- Playwright end-to-end test scripts are defined
- Extension structure
- Plugin discovery and plugin creation scripts are included
Read from README.md, package.json, lib/auth.ts, lib/fonts.ts, lib/utils.ts, lib/constants.ts, lib/api-client.ts, lib/auth-client.ts, lib/monaco-theme.ts, lib/auth-providers.ts, lib/workflow-store.ts, lib/workflow-codegen.ts, lib/workflow-logging.ts, lib/credential-fetcher.ts, lib/integrations-store.ts.
What it can do
Create automated workflows using drag-and-drop interface
User interactions with visual builder → Complete workflow automation template
Chain multiple LLM agents together
Individual AI agents and connection parameters → Sequential AI processing pipeline
Design visual workflow templates
Workflow requirements and logic → Reusable automation templates
Execute multi-step AI processing workflows
Data and workflow configuration → Processed results from chained AI operations
Configure LLM agent parameters and connections
Agent settings and workflow requirements → Configured AI workflow components
Build custom automation sequences without coding
Business logic and process requirements → Executable automation workflows
Tags
Tech Stack
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.