Span-01
Respan's behaviour classifier for AI traces that returns present, absent or not observable for each plain-language behaviour you define.
This summary is based on how curated lists describe the project; its README was not available to review.
Overview
Span-01 is a model from Respan, announced on its blog and available on OpenRouter with a Lite tier. According to the directory summary, it classifies AI traces: for each plain-language behaviour you define, such as prompt injection, hallucination or agent loops, it returns present, absent or not observable.
That three-way answer, with an explicit 'not observable' option, fits trace monitoring, where the evidence for a behaviour may simply be missing from the trace. No model card or blog text was fetched, so accuracy, latency, pricing and request format details are not available here.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Is this behaviour present in the trace? | present · absent · not observable |
How it works
- Define behaviours in plain language
- Send an AI trace to Span-01
- Receive present, absent or not observable per behaviour
- Alert or investigate based on the answers
Strengths
- Explicit 'not observable' answer
- Behaviours defined in plain language
Caveats
- Grounded only in a list summary
- Hosted model; details unknown
At a glance
- Type
- Model
- Runtime
- Hosted model (OpenRouter)
Use cases and integrations
How curated lists describe it
Respan's behavior classifier for AI traces: for each plain-language behavior you define, such as prompt injection, hallucination, or agent loops, it returns present, absent, or not observable. Span-01 and a Lite tier are on [OpenRouter](https://openrouter.ai/respan/span-01).
AnotiaWang/awesome-decision-models
Frequently asked questions
What does Span-01 decide?
Whether each user-defined behaviour is present, absent or not observable in an AI trace.
Where can I use it?
On OpenRouter, with a Lite tier also listed.
Who makes it?
Respan.
Related projects
Example of using Jev as a webhook guardrail inside the agentgateway proxy to inspect model requests and responses.
ProxyproductionRustSupervisor for unattended coding agents: one batched Jev Choice over the trajectory decides continue, pause or escalate.
ProjectbetaPythonOpenlayer library packing agent evals and guardrails (tool choice, grounding, injection, PHI) into one decision-model request.
BenchmarkbetaPythonLarge catalog of Claude Code agents, commands, hooks and MCPs, listed with community Jev mods for guardrails and subagent models.
CLIproductionJavaScriptx402 pre-payment risk-check provider combining sanctions lists, phishing feeds and simulation with Jev checks for injected instructions.
ProjectproductionCisco's security scanner for agent skills, with a System One analyzer as an advisory tier that cannot emit findings or change severity.
Agent skillproductionPython