Skip to content

See which AI providers are connected

Machine verified A machine ran this on our install and checked the result against the CLI. A test report is linked. Proven against quoxcli 0.6.0 on 2026-11-10.

Which AI can my Quox use?

One question tells you which model providers are ready, which are not, and how to add the one you want.

What you get: the providers your Quox can use right now, the ones that still need setting up, and the exact step to connect one. Nothing changes when you ask.

quox · talking to Quox

Which AI model providers are connected and ready to use on my Quox?

▤ AOCL │ L1 identity complete · allowed · 6ms

▤ AOCL │ L3 policy gate complete · proceed · 7ms

◐ QUOX │ cli ✔ QUOX │ cli │ oversized: true, 33 full size kb, 7 top keys · 8.0s

⛓ AEE │ envelope opened · ops.tool · quox_cli ok

◐ QUOX │ cli ✔ QUOX │ cli │ [ "Chat", "OpenAI", "Claude subscription", "Memory", "Qortex", "Approvals", "Receipts and verify", "Vault", "Agents", "… · 8.6s

⛓ AEE │ envelope opened · ops.tool · quox_cli ok

⛓ AEE │ envelope opened · stream_success

Two AI model providers are connected and ready on your Quox, as  provider list  prints
them:

• OpenAI (GPT models)
• Claude subscription (a signed-in Claude subscription on this server)

None of the others are connected. Each one shows as requiring setup:

• Claude API
• OpenRouter
• Hugging Face
• Groq
• Together AI
• LM Studio
• vLLM
• Ollama
• Custom OpenAI-compatible endpoint

Ollama and the custom endpoint did not appear among the ready entries. I did not read
their individual status lines.

To add one, an org admin runs  /connect <provider> . For example,  /connect anthropic 
connects the Claude API, and  /connect openrouter  connects OpenRouter.

31.2s · 2 tools · 0 approvals

A real turn on our own install (CLI 0.6.0, 9 Nov 2026), replayed at speed. Names changed. Type the same words into quox.
✔ Machine verified against the CLI on 9 Nov 2026 · test report mw9w3l

Quox read the real provider list for your org, the same one the terminal shows below. It named the two that are ready, listed every provider that still needs setting up, and said plainly which status lines it did not read. Connecting a provider is an org admin step (/connect <provider>), so Quox told you how rather than doing it.

provider list
quox provider listConnections for this org:● OpenAI  ready on this server (an instance-wide key, not in your org vault)○ Claude API  not connected · /connect anthropic○ Ollama  not connected · /connect ollama http://<host>:11434○ OpenRouter  not connected · /connect openrouter○ Hugging Face  not connected · /connect huggingface○ Groq  not connected · /connect groq○ Together  not connected · /connect together○ LM Studio  not connected · /connect lmstudio http://<host>:1234○ vLLM  not connected · /connect vllm http://<host>:8000○ Custom  not connected · /connect custom https://<host>/v1● Claude (subscription, dev only)  ready on this server · /model claude-cli1 other integration in your vault: quox provider list --allAn instance-wide key set by your admin can also serve chat; it is not visible from here.
From the same scratch workspace, read after the turn above.

A filled dot is ready, an empty one is not connected, and each line ends with the step that connects it. quox provider list --all also shows the other integrations in your org vault.

Quox and the terminal read the same provider list for your org. Reading changes nothing. Connecting or removing a provider is a separate step an org admin takes.