safer-with-jev
Public Neon Function gateway that has Jev judge request bodies for prompt injection or unsafe replies before optionally forwarding them.
Overview
Safer with Jev is a public showcase of TypeSafe Jev judgments, with a newspaper-style site on Vercel and an API running as a Neon Function. The gateway inspects a request body with Jev and, if it passes, optionally forwards the same bytes to a caller-chosen HTTPS URL given as a target parameter. Endpoints include block-prompt-injections, block-unsafe-replies, ask-jev for a free-form yes/no question about a text, and a nice-try demo.
Without a target, the call returns a 200 judgment. With a target, a pass forwards the request, while review and block verdicts return 403 and never forward; response headers report the action and Jev latency. There is no Safer API key: only the server holds TYPESAFE_API_KEY. Agents can read the site as Markdown or via AGENTS.md and llms.txt. It is deployed with Bun and the Neon CLI.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Does this request contain a prompt injection: pass, review or block? | pass · review · block |
| Choice | Is this reply unsafe: pass, review or block? | pass · review · block |
| Yes / no | Is this good text? (arbitrary ask-jev question) | — |
How it works
- POST the body to an inspection path, optionally with a target URL
- Jev judges the content
- Pass forwards the same bytes; review or block returns 403
- Read x-neon-action and Jev timing headers
Strengths
- No client key needed for the public demo
- Forwards bytes unchanged on pass
- Clear review/block semantics
Caveats
- Public showcase; thresholds and questions not documented
- Routes traffic through a third-party demo service
Getting started
curl -i "https://api.safer-with-jev.com/block-prompt-injections" -H "Content-Type: text/plain" --data-binary 'text'; to self-host: bun install && neon deploy --env .env.local.At a glance
- Type
- Proxy
- Status
- prototype
- Language
- TypeScript
- Runtime
- Neon Function + Next.js site
- Has code
- yes
- Has tests
- yes
Use cases and integrations
How curated lists describe it
Cloud infrastructure: Neon Function proxy for the Neon AI Gateway that routes decisions with Jev.
yibie/awesome-jevAn HTTP inspection gateway that checks request content with Jev before optionally forwarding it to an HTTPS destination.
logicrw/awesome-jev-projectsNeon Function proxy for the Neon AI Gateway with TypeSafe Jev routing.
hellogumbo/awesome-jevNeon Function proxy for the Neon AI Gateway. Jev classifies each request and routes it to the right downstream model.
MrJev/awesome-jev
Frequently asked questions
What does Jev decide?
Whether a request contains a prompt injection or a reply is unsafe, returning pass, review or block.
Do I need an API key?
Not for the public demo; the server holds the TypeSafe key.
What happens on block?
The request is not forwarded and the gateway returns 403.
Related projects
Large catalog of Claude Code agents, commands, hooks and MCPs, listed with community Jev mods for guardrails and subagent models.
CLIproductionJavaScriptMCP proxy and hooks that screen agent tool calls before they run and their results before the agent reads them, using Jev judgments.
MCP serverbetaTypeScriptXano backend template where Jev judges each agent tool call as safe, needs review or dangerous, and uncertain calls wait for a human.
ProjectbetaXanoScriptEffect-based safety gate for coding-agent shell commands: fast rules first, then Jev decides allow/deny plus eleven risk checks.
ProjectbetaTypeScriptDeepSeek Harness plugin where Jev classifies risk, irreversibility, task fit and injection for every tool call before it runs.
PluginbetaTypeScriptOpen-source agent guardrails that use Jev or LAYA to check prompts, retrieved content, tool calls and outputs for injection and risk.
ProjectbetaPython