Ollama · Qwen · DeepSeek · Llama · Local bridge
DNSai with Ollama and local models
Keep the model on your own hardware and still answer questions about your domains. Ollama has no built-in MCP client, so either launch an MCP-capable agent from Ollama, use a local chat app with MCP (LM Studio, Open WebUI), or run DNSai's small bridge script.
What you can ask Ollama and local models
Example answers use made-up domains.
“Summarise this week's DNS changes on my favourite domains.”
Connect in 3 steps (about 5 minutes)
- Create a DNSai API key.
- Pick a path: ollama launch an MCP agent, a local chat app with MCP, or the DNSai bridge script.
- Ask your local model about your domains.
Signed-in Enterprise customers get the exact click-by-click steps and ready-to-paste configuration for Ollama and local models, a one-click API key for it, and a live connection check.
Ollama and local models requirements: Ollama (or vLLM / any OpenAI-compatible server) and a model that supports tool calling, such as Qwen 3, DeepSeek, Llama 3.1+ or Hermes. DNSai Enterprise plan.
Good for
- Air-gapped analysis where only DNSai's answers leave the network
- Cheap high-volume triage of DNS changes
Security
- The bridge runs on your machine and only calls DNSai's read-only endpoints with your key.
- Account and organization owners can make API keys and assistants read-only, so they can read your data but never change it or start scans.
- Starting a credit-spending Deep Scan always needs an estimate and your explicit confirmation.
- DNSai data comes from public DNS and web records. Your watchlists, notes and tags stay private to your account and are only shared with an AI tool you connect yourself.
Connect Ollama and local models to DNSai
The API and AI connectors are included with DNSai Enterprise plans. Every member of your organization shares one set of limits, and every call counts from the first one.
Other integrations
Last checked 2026-10-29.