Score8.0
Popularity151.0
Risklow
TierGold
Score breakdown
Usefulness8.4
Novelty9.1
Momentum10.0
Maturity8.8
Open-source/build7.4
Evidence7.2
Workflow potential9.1
Setup ease6.5
Popularity is tracked separately. Support, ads, sponsorships, and tips never affect these signals.
Why it matters
Useful for RL researchers and agent-training teams that need a production-grade agentic RL training framework that supports both whitebox (Python tool loops) and blackbox (HTTP agents like opencode / openclaw) through a unified interface; for teams that need TITO (Token-In-Token-Out) to avoid retokenization drift by encoding only each turn's append delta and splicing token IDs incrementally - the
Who should use it
Who should skip it
Skip if the source link, docs, or setup requirements do not match your workflow.
Risk explanation
Risk label needs manual review.