Generate a data analysis pipeline for me
Quick Start
One command. 60 seconds. Done.
From global install to your first running pipeline, TaskMeld auto-detects your environment, provisions the Agent runtime, and wires up the Gateway — so you focus on workflows, not infrastructure.
npm install -g taskmeld#01
One Command, Globally Ready
No Docker, no extra dependencies. Global install completes in under 30 seconds, ready to use immediately.
#02
Auto-Detect, Smart Provision
Scans your system, installs the OpenClaw Agent runtime on demand, and configures the Gateway — all in one step.
taskmeld init#03
Launch & Build, Zero Wait
Spins up the Agent runtime, Web Console, and pipeline engine together — open the browser and start orchestrating.
taskmeld server startBuilt for Real Work
Intelligent Main Agent
Describe what you need — the agent generates pipelines, schedules execution, and archives artifacts. One conversational entry point to run the entire platform.
Wevra Agent · Session14:32
Send
Visual Pipeline Canvas
Drag-and-drop DAG orchestration with parallel groups and branch routing. Pipelines created by the agent render in real time — fine-tune by hand whenever you want.
Pipeline · Data Analysis
ReadyData CollectionWaiting
@collector
Cleaning & TransformationWaiting
@cleaner
Analysis & ComputationWaiting
@analyst
Report GenerationWaiting
@reporter
Runtime Logs
@collector Collected 12,340 raw records
@cleaner Cleaned, 2,341 duplicates removed, 9,999 remaining
@analyst Analysis complete, 15 metrics generated
@reporter Report generated: data_report_202405.pdf
Artifact Delivery & Pipeline Chaining
Pipeline A delivers artifacts to Pipeline B, auto-triggering on validation. Multiple pipelines feed the same target with built-in queuing, sequential consumption, and guaranteed delivery.
A1Ready
A2Ready
A3Ready
Buffer0
A1
A2
A3
BAwaiting Delivery
Local-First Deployment
Run it on your own machine. Install, execute, and store artifacts locally — private, secure, and fully under your control.
Local HostingData stays on device
One-Click InstallINSTALL
Start quickly without remote-side configurationLocal RuntimeLOCAL
Pipelines, agents, and schedules run in your own environmentLocal Artifact StorageSTORAGE
Logs, results, and files are written to local directories for audit and migrationEnd-to-End Observability
WebSocket real-time push, timeline logs, and artifact tracing. Every step is replayable — no more blind troubleshooting.
Runtime LogsLIVE
14:32:01[Agent α] Starting data collection...
14:32:03[Agent β] Cleaning & transformation launched in parallel
14:32:05[Agent α] Collection complete · 12,340 records
14:32:06[Main Agent] Branch routing → 3 downstream nodes
14:32:08[Agent β] Cleaning complete · 2,341 duplicates removed
14:32:11[Agent γ] Analysis report generated ✓
FAQ
How does TaskMeld relate to OpenClaw?
Three-layer architecture: OpenClaw executes (each pipeline node runs an Agent), TaskMeld orchestrates (DAG dependencies, scheduling, artifact tracking), and Wevra Agent intelligently directs (manage the entire platform via natural language). In short — OpenClaw runs agents, TaskMeld orchestrates them.
How is it different from Jenkins / Airflow / n8n?
They're general-purpose workflow engines; TaskMeld is purpose-built for agent pipelines. Key differences: the main agent auto-orchestrates without manual wiring, artifact tracking and full-chain log replay are built in, and each node binds an OpenClaw Agent for actual execution — not scripts or containers.
What are the system requirements?
Node.js 18 or higher. The OpenClaw Agent runtime is auto-detected and installed during taskmeld init, or you can install it manually or connect to a remote Gateway. Full support is currently available on Windows; macOS and Linux support is in progress.
What happens when an agent errors out?
Pipelines support per-node automatic retry (configurable policies) and branch fallback. Full-chain timeline logs record every input and output, so you can pinpoint failures to the exact node and agent. You can manually retry failed nodes via the Web Console or CLI.
What can Wevra Agent do?
It ships with 40 built-in tools covering pipeline CRUD, agent lifecycle management, system monitoring, memory, and skills. You can create pipelines, check run status, diagnose failures, and even let it auto-optimize execution strategies — all through natural language. Supports DeepSeek, OpenAI, and other LLM providers.
Is it free?
Completely free, MIT open source. The project is community-driven — contributions, issues, and documentation improvements are welcome.