Backend\'ler

Nexvexa Agent bes farkli LLM saglayicisini destekler ve aralarinda anlik gecis yapabilirsin.

Desteklenen saglayicilar

Saglayici	Komut	Notlar
OpenAI	`/provider openai`	GPT-4o, GPT-4o-mini, GPT-4.1, o-series
Anthropic	`/provider anthropic`	Claude 3.5 / 3.7 / 4 Sonnet/Opus/Haiku
Google	`/provider google`	Gemini 1.5 / 2.0 Pro/Flash
Ollama	`/provider ollama`	Tamamen yerel; llama, qwen, mistral, deepseek vb.
Custom (OpenAI uyumlu)	`/provider custom`	Groq, Together, OpenRouter, kendi proxy\'in

Custom endpoint ekleme

/custom add groq https://api.groq.com/openai/v1 llama-3.3-70b-versatile gsk_...
/custom use groq

Ollama hizli kurulum

# Sistem geneli (Linux/macOS)
curl -fsSL https://ollama.com/install.sh | sh
ollama pull qwen2.5-coder:7b
nexvexa-agent
/provider ollama
/model qwen2.5-coder:7b

Backend modu

`/backend direct`	Saglayici API\'sina dogrudan baglan (varsayilan).
`/backend local`	Yerel TinyGPT yonlendirici (deneysel).

Test

/test
# baglanti basarili / model yaniti / latency

Hata durumlarinda

HTTP 401 — API key gecersiz veya hesabin pasif. /keys ile kontrol et.
HTTP 429 — Kotayi astin. Modeli degistir (/model gpt-4o-mini) veya butce duzenle.
fetch failed — Network izinleri (sandbox), proxy veya cografi engel.