Respan launches Guardrails, claims top average F1 across 8 safety categories, free with its gateway
ycombinator · x · 2026-09-29
Respan introduced Guardrails, built on its Span-01 model, to catch prompt injections, jailbreaks, and unsafe content before requests reach your models and tools.
The company claims the highest average F1 of any guardrail it benchmarked across 8 categories, and it ships free and natively with the Respan gateway. Relevant for developers building AI security layers.
More from coding & agent
- OpenRouter coding model share: GLM 5.3 Flash leads at 29.2%, DeepSeek V4.1 Flash at 25.6% — togethercompute · 2026-09-29
- InstaCloud raises $8M seed for agent-native serverless cloud that deploys itself — Shruti_0810 · 2026-09-29
- Routing with small models is 200x faster and 400x cheaper than LLMs, reshaping AI software architecture — Pavan_Belagatti · 2026-09-29
- SRE copilot built on Hindsight memory stops agents from repeating past incidents — sathvik1233 · 2026-09-29
- VeilMind prototype isolates private AI memories while sharing generalized lessons — Logical_Leading8882 · 2026-09-29
- Open-source Manus-style skill recreates CellMotion animations, fully editable — yihui_indie · 2026-09-29