Check your server
Choose n8n
This step only looks. It does not run commands inside n8n.
The sandbox joins this network. n8n stays unchanged.
Check whether this n8n already includes instance-ai.
Required n8n prerequisite. Set N8N_ENABLED_MODULES=instance-ai before Relmio can continue.
How to add it
If other modules are already set, append instance-ai as a separate comma-delimited entry and keep the existing ones. Do not replace unknown values.
- Update n8n with your existing deployment workflow.
- Redeploy or restart n8n, then check that it is healthy.
- Reconnect and run discovery again.
Relmio cannot apply this change. It will not edit the n8n Compose file, image or environment, restart or recreate n8n, or exec into n8n. It never shows the raw setting.
Review the plan
Nothing is written until you approve this plan.
The wizard will
- Create
/docker/n8n-openai-oauth/assistant-sandbox - Start only the managed AI Assistant companion project
- Attach sandbox API to Not selected
- AI Assistant prerequisite has not been reviewed
- Web search selection has not been reviewed
- Generate separate 256-bit sandbox, runner, and private SearXNG secrets
- Check sandbox health, expected services, and no published host port
The wizard will not
- Edit, rebuild, stop, restart, recreate, or exec into n8n
- Upload your OpenAI Platform API key or ChatGPT/Codex OAuth
- Publish sandbox, runner, or SearXNG ports to the SSH host
- Enable a model provider or change n8n settings automatically
AI Assistant companion is ready
Use these values in n8n. The sandbox key appears only in this result.
Preview with human review. Review every generated workflow. Use Daytona for production.
1. Sandbox settings
- Sandbox Service URL
Not installed- Sandbox API key
Shown after installation
2. Web search settings
- SearXNG Instance URL
Web search disabled
Web search is off, so no SearXNG service or settings file was installed. Its secret stays private.
3. Model provider
- Model provider
OpenAI- Model
Keep the current supported selection
Enter your OpenAI Platform API key in n8n. Relmio does not read, receive, or store that key.
Keep your current supported n8n model, including openai/gpt-5.6-sol, while n8n accepts it. Change it only if n8n rejects it or you choose another supported model. A listed model is not proof that every workflow works.
No setup can promise an account will never be flagged.
Platform account limits
Use a dedicated OpenAI Platform project and key. Keep it in n8n, set spend limits, review use, and do not expose this path publicly. Relmio cannot add per-user safety identifiers or moderation to n8n’s direct provider path.
Optional Ollama-style custom endpoint
n8n’s custom screen defaults to an Ollama-style endpoint. Use it only with a separate n8n-reachable Relmio endpoint that uses your own OpenAI Platform API key. This wizard does not create one.
- Base URL
your Platform-key-backed Relmio /v1 endpoint- API key
- Use that endpoint’s private Relmio client credential. It is not an OpenAI-issued API key.
- Model ID
- Use a model ID exposed by that Platform-key-backed endpoint.
The existing ChatGPT/Codex OAuth sidecar may work, but it remains unofficial, private, and policy-uncertain. It is not auto-selected, enabled, or presented as TOS-approved.