Managed Apps
Ready-to-use AI tools — chat interfaces, workflow builders, agent platforms — provisioned and hosted by Servada on a per-org basis, already connected to the models configured for you.
What you get
Managed Apps lets your team use popular open-source AI tools without running any infrastructure. From the Apps page inside your organisation, you pick an app from the catalogue and click Provision; a few moments later Servada spins up a dedicated instance for your org — on its own subdomain, with its own admin login, already pointed at the models configured for you.
Hosting, updates, credentials, and wiring to the models are handled by Servada. Your team focuses on building and using the app.
What you can spin up
The apps below are available today. More will be added over time.
A clean, ChatGPT-style interface for your team.
Give your colleagues a polished chat experience against the models configured for your org. Users sign in, pick a model, and chat — with conversation history, file uploads for Q&A, and prompt libraries. The starting point for teams that want "ChatGPT, but routed through the gateway."
Build chatbots and AI workflows visually, no code.
Drag-and-drop builder for chatbots, retrieval-augmented search, and multi-step AI flows. Great for teams who want to prototype AI features quickly without writing code — connect a prompt to a document store to a model and ship.
Automate everything, with AI as one step in the chain.
General-purpose automation platform with hundreds of integrations — databases, email, Slack, APIs, and more. Pick this when AI is one piece of a bigger workflow (e.g. summarise incoming tickets, then route them). Strong fit for ops teams.
Getting in
Once an app shows as running, click Open to launch it in a new tab, and Reveal credentialsto get the admin login on first visit. Each app has its own built-in account system — once you're in, you add your team, manage permissions, and use the app exactly as its documentation describes.
What's pre-wired in the app
When you provision with Auto-create a token, your app boots with the gateway already wired in:
- OpenWebUI — the gateway is the default OpenAI connection. Pick a model and chat.
- n8n — an
Inferada (default)OpenAI credential is created for the admin user. Both classic OpenAI nodes and the AI / LangChain nodes share this one credential type. Pick it on any node, choose a model, run. - Flowise — same
Inferada (default)credential is created, and the container'sOPENAI_BASE_URLdefaults to the gateway. Use theOpenAI Custom Model(orOpenAI Embeddings Custom) node, pick the credential, type a model id — leave Base Path blank.
If the credential is missing for some reason (a transient seed failure during provisioning), you can always add it manually with the same name and the same inf_… token — the rest of the wiring still works.
Who owns the auto-created token
When you provision an app with Auto-create a token, Servada mints a fresh Inferada API token so the app can talk to the gateway out of the box. That token row belongs to a real user — pick one from the Token owner dropdown in the provision dialog. The owner sees the token in their personal API tokens list inside the org.
Pick a member of thisorganisation, not yourself if you're a super-admin working from a different org — otherwise the token shows up in your token list rather than the org's, which makes it harder for the org to audit and rotate. When the app is destroyed, the token is revoked automatically.