AEVNI Public API

Optimize and create prompts programmatically from any app or pipeline.

The AEVNI API exposes two operations:

• Optimize — reduce the token count of an existing prompt while preserving intent.
• Create — generate a well-structured prompt from a plain-language description.

Both endpoints are authenticated and require an X-AEVNI-Key header. You also bring your own LLM API key via X-Provider-Key.

⚡ Playground

Probá la API en vivo con tu propia AEVNI key. Las keys quedan solo en tu navegador.

Authentication

Every request must include two headers:

Alternatively, pass api_key in the JSON body instead of X-Provider-Key.

Copy# Example — curl
curl -X POST https://aevni.website/api/v1/optimize \
  -H "Content-Type: application/json" \
  -H "X-AEVNI-Key: aevni_your_key_here" \
  -H "X-Provider-Key: AIza..." \
  -d '{"prompt": "Explain quantum entanglement", "provider": "gemini", "mode": "pro"}'

Errors

All errors return a JSON body with error (human-readable message) and code (machine-readable slug).

{
  "error": "API key inválida o revocada.",
  "code":  "invalid_key"
}

Rate limits

Both v1 endpoints allow 60 requests per minute per IP. Beyond that you'll receive HTTP 429. Provider-side rate limits (Gemini RPM, etc.) are separate — the rate_limited code distinguishes them.

POST /v1/optimize

Reduce the token count of an existing prompt while preserving its meaning, structure, and intent.

Request body

Response

Copy{
  "optimized":      "...",          // optimized prompt in original language
  "optimized_en":   "...",          // English version (often saves more tokens)
  "detected_lang":  "es",          // "es" or "en"
  "prompt_type":    "task_structured",
  "report": {
    "original_tokens":      120,
    "optimized_tokens":     74,
    "optimized_en_tokens":  68,
    "tokens_saved":         46,
    "token_reduction_pct":  38.3,
    "chars_saved":          210,
    "density_original":     4.2,
    "density_optimized":    5.1,
    "detected_language":    "es"
  },
  "recommendation": {
    "model":       "gemini-2.5-flash",
    "reason":      "...",
    "cost_tier":   "low"
  },
  "meta": {
    "provider":      "gemini",
    "mode":          "pro",
    "mode_resolved": "pro"
  }
}

Example

Copycurl -X POST https://aevni.website/api/v1/optimize \
  -H "Content-Type: application/json" \
  -H "X-AEVNI-Key: aevni_..." \
  -H "X-Provider-Key: AIza..." \
  -d '{"prompt":"You are an expert assistant. Your job is to help users with their questions in a kind and friendly manner. Always be polite.","provider":"gemini","mode":"auto"}'

POST /v1/create

Generate a well-structured prompt from a plain-language description of what you need.

Request body

Response

Copy{
  "prompt_type":        "task_structured",
  "created_prompt_es":  "...",   // prompt in user's language
  "created_prompt_en":  "...",   // English version
  "tokens_es":          95,
  "tokens_en":          88,
  "explanation":        "Why this structure was chosen...",
  "tips":               ["Add examples to improve accuracy", "..."],
  "meta": {
    "provider":    "pollinations",
    "tone":        "technical",
    "audience":   "developer",
    "forced_type": null
  }
}

Example

Copycurl -X POST https://aevni.website/api/v1/create \
  -H "Content-Type: application/json" \
  -H "X-AEVNI-Key: aevni_..." \
  -d '{"description":"A customer support agent for a SaaS app","tone":"formal","audience":"general","provider":"pollinations"}'

Providers

Pass the provider string in your request body.

Modes

Prompt types

AEVNI detects the structural type of your prompt automatically. You can override it with forced_type in /v1/create.