Security & Guardrails11.4ms Reflex$0 / MIT License
38,900 active downloads
JSON Schema Repair Reflex
Edge schema fixer and bracket repair for malformed LLM outputs
jevproxy // json-repair-reflex (reflex kernel)
$npx jevproxy run json-repair
JevProxy Intercept Resolved(Mode: DIRECT_REFLEX)
11.4ms|Cost: $0.0001|0 reasoning tokens burned
STDOUT • Tool Call Output:Status: 200 OK
agent.dispatch(repair_json_payload)
payload: { "raw_json": { "type": "string", "description": "Truncated or malformed JSON string" } }...
✔ Decision short-circuited in 11.4ms without roundtrip to frontier LLM.
Upstream token bill saved: $0.0150 on this turn.
JSON Schema Repair Reflex
Compatible with Cursor, Claude Code, Windsurf, OpenCode
TRADITIONAL LLM CALL:UNOPTIMIZED
• Median Latency: 1420 ms
• Cost per Turn: $0.0150
• Mode: Full KV Cache Reload & TTFT Prefill
JEVPROXY REFLEX KERNEL:77x FASTER
• Median Latency: 11.4 ms
• Cost per Turn: $0.0001
• Accuracy: 99.99% deterministic
1-Click CLI Execution
Run this tool accelerated through the JevProxy gateway without manual wiring:
npx jevproxy run json-repairOpenAI / Anthropic Tool Schema
JSON SpecificationPaste this schema into your agent tools definition or Cursor extensions:
{
"type": "function",
"function": {
"name": "repair_json_payload",
"description": "Repair malformed JSON output from autoregressive LLMs",
"parameters": {
"type": "object",
"properties": {
"raw_json": {
"type": "string",
"description": "Truncated or malformed JSON string"
}
},
"required": [
"raw_json"
]
}
}
}Technical Architecture & Usage
Zero-Roundtrip JSON Recovery
When a model outputs truncated JSON, traditional frameworks spend 1,400ms asking the model to fix itself. JSON Repair Reflex fixes it in 11.4ms.
Accelerate JSON Schema Repair Reflex with JevProxy
Get 5,000,000 free decision tokens. Eliminate the 3-second tool freeze in Cursor and Claude Code in under 60 seconds.