Weavr

Starting the Server

Launch the Weavr server to access the web interface and API.

node weavr.mjs serve

Open http://localhost:3847 in your browser.

Build first: Make sure you've run npm run build:web before starting the server.

Dashboard

The dashboard provides a comprehensive overview of your Weavr instance:

Weavr Dashboard

Key Features

  • AI Model - Shows your configured AI provider, model, and authentication status
  • Token Usage - Tracks input/output tokens and API requests across all AI operations
  • Workflows - Displays active, paused, and total workflow counts with success rate
  • Active Workflows - Lists deployed workflows with status, triggers, and quick actions (Run Now, Pause, Resume)
  • Recent Runs - Shows latest workflow executions with status and duration
  • Quick Actions - One-click access to create workflows, view all workflows, or access settings
Real-time updates: The dashboard automatically refreshes when workflows run, showing live status via WebSocket connection.

Workflows

The Workflows page lets you manage your workflow definitions:

Viewing Workflows

Browse all workflows loaded from your workflows directory. Click a workflow to see its YAML definition and configuration.

Creating Workflows

Create new workflows directly in the web UI with the built-in YAML editor.

Running Workflows

Execute any workflow with optional input data. The UI shows real-time execution progress.

AI Workflow Assistant

The Workflow Builder includes an AI-powered assistant that helps you create workflows by describing what you want in plain English.

How to Use

  1. Open the Workflow Builder (create new or edit existing workflow)
  2. Click the AI Assistant button in the toolbar
  3. Describe the workflow you want to create
  4. The assistant generates YAML that you can review and apply

Example Prompts

# Create a new workflow
"Create a workflow that monitors a GitHub repo for new PRs
and posts a summary to Slack"

# Add steps to existing workflow
"Add an AI agent step that analyzes the PR diff
before posting to Slack"

# Use MCP tools
"Create a workflow that uses the filesystem MCP to
backup important files daily"

Available Tools

The assistant is aware of all available tools including:

  • Built-in tools - web_search, http_request, read_file, write_file, etc.
  • MCP tools - Any tools from enabled MCP servers (filesystem, git, database, etc.)
Tip: The assistant knows about your enabled MCP servers. Enable the tools you need in Settings before asking the assistant to use them.

Settings

Configure your Weavr instance from the web UI.

AI Configuration

  • Provider - Choose between Anthropic, OpenAI, Google, and 20+ other providers
  • Model - Select from available models for your provider
  • API Keys - Configure API keys (stored in local config)

MCP Server Management

Enable and disable MCP servers in real-time without restarting the gateway. The Settings page shows a catalog of 20+ pre-configured MCP servers:

Category Servers
Filesystem filesystem, everything-search
Git & Code git, github, gitlab
Database postgres, sqlite, mysql
Browser puppeteer, playwright, browserbase
Search brave-search, exa
Cloud aws-kb-retrieval, cloudflare
Productivity google-drive, slack, linear, obsidian
Dev Tools sentry, sequential-thinking

Dynamic Enable/Disable

Toggle MCP servers on/off directly from the UI. Changes take effect immediately:

  • Enable - Server starts and tools become available to agents
  • Disable - Server stops and tools are removed
  • Status - Shows "running" or "not connected" in real-time
No restart required: MCP servers can be enabled/disabled while workflows are running. New workflow runs will use the updated tool set.

Run History

View the execution history of all workflows:

  • Status - Success, failed, or running
  • Duration - How long the run took
  • Step details - Individual step outputs and errors
  • Logs - Full execution logs for debugging

Chat Interface

Interact with an AI agent directly from the web UI. The chat interface supports multi-turn conversations and can execute workflows on your behalf.

Requires AI provider: Configure an Anthropic or OpenAI API key to use the chat feature.

REST API

The server exposes a REST API for programmatic access.

Endpoints

Method Endpoint Description
GET /api/workflows List all workflows
POST /api/workflows Create a workflow
GET /api/workflows/:name Get workflow details
POST /api/workflows/:name/run Execute a workflow
GET /api/runs List all runs
GET /api/runs/:id Get run details
GET /api/config/status Check configuration status
GET /api/stats Get dashboard stats (AI config, token usage, workflow counts)

Examples

List Workflows

curl http://localhost:3847/api/workflows

Run a Workflow

curl -X POST http://localhost:3847/api/workflows/hello-world/run \
  -H "Content-Type: application/json" \
  -d '{}'

Run with Input Data

curl -X POST http://localhost:3847/api/workflows/process-data/run \
  -H "Content-Type: application/json" \
  -d '{"input": {"file": "data.csv"}}'

Create a Workflow

curl -X POST http://localhost:3847/api/workflows \
  -H "Content-Type: application/json" \
  -d '{
    "name": "my-workflow",
    "yaml": "name: my-workflow\nsteps:\n  - id: greet\n    action: log\n    with:\n      message: Hello!"
  }'

Check Run Status

curl http://localhost:3847/api/runs/run_abc123