OpenClaw Trace Contributor — AI Agent Data

About The Role

Get paid for OpenClaw sessions you've already completed. We're collecting structured exports of real sessions with 150+ turns of substantive exchange, real tool use, and meaningful progress toward a coherent objective.

This is not a generation task. We are not asking you to run new sessions, synthesize traces, or pad turn counts to qualify. We want exports of work you've already done.

Fully remote, fully asynchronous, paid per accepted submission.

• Organization: Alignerr
• Type: Per-Submission Contract
• Location: Remote
• Commitment: Flexible — submit one trace or many
  
  

What You'll Do

• Identify a completed OpenClaw session with more than 150 turns and meaningful progress toward a coherent objective
• Run the provided export prompt to generate a structured trace of the session
• Bundle any referenced files into an artifacts/ directory alongside the trace
• Redact PII, credentials, network identifiers, and business or customer data before submitting
• Package the trace (and artifacts, if any) and complete the ontology fields in the labeling interface
• Confirm rights and compliance — no employer infrastructure, no confidential data, no customer or client work
  
  

Who You Are

• Active or recent OpenClaw user with real sessions exceeding 150 turns
• Experienced with long-horizon agentic work involving real tool use: code execution, file edits, browser actions, or API calls
• Comfortable redacting PII, credentials, and proprietary content without breaking technical coherence
• Clear on your rights to share — personal work, open-source contributions, or sessions with explicit permission
• Detail-oriented and self-directed — able to export, review, and submit without hand-holding
  
  

Nice to Have

• Sessions in the 301–500 or >500 turn buckets
• High-complexity traces with long planning horizons, varied tools, and recovery from compound failures
• Sessions involving browser DOM captures, screenshots, skill files, or system prompts
• Experience with redaction in cyber security, DevOps, or other identifier-heavy domains
• Multiple distinct sessions across different objectives or model versions
  
  

Why Join Us

• Get paid for agentic work you've already done — no new sessions, no synthesis, no fabrication
• Fully remote and asynchronous — export and submit on your schedule, no meetings
• Flat per-submission rate with no cap on the number of qualifying traces you can contribute
• Contribute to a real dataset of long-horizon, tool-rich agent behavior — the kind that's hard to source authentically
• Resubmission allowed when issues are fixable; rejected traces come back with a brief reason
• Build a track record on a project where authenticity and quality are rewarded directly