Providers

Bring your own AI for sitemap assistants. Use managed AI for RAG.

Basic website assistants can run on your provider account. Document upload, vector retrieval, and advanced plans run on Askly managed inference and embeddings so production retrieval stays consistent.

Bring your own key

Your provider. Your terms.

Already pay for a model you trust? Connect it on Sitemap Assistant. Askly handles hosting, sitemap crawling, retrieval, and design while you keep the basic model relationship.

Available on the Sitemap Assistant plan for website and sitemap assistants.
Keep spend and data terms under your own provider contracts.
Keys encrypted with KMS, scoped to your site, tested before they go live.
Upgrade to Knowledge RAG when you need managed document embeddings.

Managed AI

Or let Askly handle it.

No accounts, no keys, no decisions. Turn it on and Askly routes the right model for the job, with sensible defaults and graceful fallback.

The fastest path — no provider account to set up.
Askly routes inference and bills simple usage credits.
Routing policies: cheapest, fastest, best quality, long context.
Automatic fallback if a provider has an outage.

BYOK keys stay locked away server-side and never reach a browser. RAG plans use Askly managed inference plus managed Cloudflare embeddings.

The matrix

Every major model. One gateway.

Twenty providers, normalized to one request and response shape. A single OpenAI-compatible adapter covers most of them — with native adapters where it matters.

OpenAI
Native + compatible
Anthropic
Native Messages
Google Gemini
Native
Cohere
Rerank
Mistral AI
OpenAI-compatible
xAI (Grok)
OpenAI-compatible
Azure OpenAI / Foundry
Native
AWS Bedrock
Converse
Google Vertex AI
Native
Cloudflare Workers AI
Workers AI
Groq
OpenAI-compatible
Cerebras
OpenAI-compatible
SambaNova
OpenAI-compatible
OpenRouter
OpenAI-compatible
Together AI
OpenAI-compatible
Fireworks AI
OpenAI-compatible
DeepInfra
OpenAI-compatible
Novita AI
OpenAI-compatible
Hugging Face
OpenAI-compatible
Perplexity
OpenAI-compatible
Custom OpenAI-compatible (public endpoint)
OpenAI-compatible
DeepSeek
OpenAI-compatible
Alibaba Qwen (DashScope)
OpenAI-compatible
Moonshot AI (Kimi)
OpenAI-compatible
Zhipu AI (GLM)
OpenAI-compatible
MiniMax
OpenAI-compatible
Baichuan
OpenAI-compatible
01.AI (Yi)
OpenAI-compatible
StepFun
OpenAI-compatible
Xiaomi (MiMo)
OpenAI-compatible

Native: OpenAI · Anthropic · Gemini · Azure · Bedrock · Cloudflare · Cohere rerank. Everything else through the compatible adapter.

Pick your model for sitemap. Let us run RAG.

Connect a key for a basic sitemap assistant, or upgrade to managed inference and embeddings for documents.

Start free BYOK docs