Workspace Setup For Admins
Streamlined checklist for preparing a workspace before end users start chatting.
Use this page as a practical runbook for preparing a new OrcAI workspace.
Outcome
After completing this checklist, users should be able to:
- Sign in and join the correct organisation
- Start chats with an approved model
- Use published bots with ready content collections
- Operate within defined quota and access boundaries
1. Confirm Access And Organisation Context
- Sign in with an admin account.
- Ensure your active organisation is correct.
- Verify you can see Administration in the sidebar.
Missing Administration?
If Administration is missing, your account likely does not have the required capabilities.
2. Add Or Validate A Provider
Go to Administration > Providers.
- Add provider entries for the inference backends your workspace should use.
- Validate connectivity and credentials according to your deployment setup.
TODO
Add provider-by-provider field reference once provider forms are fully documented.
3. Add A Specific Model
Go to Administration > Models.
- Click Add Model.
- Select the provider you want the model to use.
- Register the exact model identifier required by that provider.
- Save, then confirm the model appears in the list.
Recommended validation:
- Open a new chat.
- Select the newly added model.
- Send a simple test prompt and confirm a successful response.
4. Configure Quota Pools
Go to Administration > Quotas.
- Create quota pools aligned with your cost and usage policy.
- Assign pools according to your organisation policy.
In chat, users can see remaining pool amounts when quota badge data is available.
TODO
Add quota assignment examples for class, team, and pilot environments.
5. Prepare Content And Retrieval
Go to Library > Content.
- Add representative source material.
- Confirm processing completes successfully.
Then go to Library > Repositories:
- Create or edit repository blocks.
- Attach ready content items.
- Keep repository block status as Ready for production use.
Use the repository playground to sanity-check retrieval quality.
6. Prepare Bots
Go to Library > Bots.
- Create/edit bots with the guided setup.
- Attach one ready behaviour block.
- Attach ready repository blocks with content.
- Configure sharing so the intended user groups can use the bot.
- Publish the bot.
7. Set Up Users, Groups, And Invitations
Go to Administration > Users and Administration > Groups.
- Create or verify groups.
- Invite users.
- Place users in the right groups.
- Validate that access to bots/content matches expectations.
8. Final Smoke Test
Run this short test from a non-admin user account:
- Accept invitation and select organisation.
- Start a chat using a published bot.
- Confirm model selection works.
- Confirm retrieval-backed answers use expected content.
- Confirm permissions block access to resources outside the user scope.
Self-Hosting Note: Custom Embedding And Image Processing Models
Custom embedding model and image-processing model definitions are supported only in self-hosted deployments.
For managed/shared deployments, treat these as platform-level settings controlled by the hosting operator.
If you self-host and change embedding model or dimensions on an existing indexed dataset, you must recreate the vector collection and reprocess content.
TODO
Add explicit configuration variable mapping for embedding and image-processing model configuration once central config docs are finalised.