Skip to main content

Quickstart with FlowerDocs

This guide adds the Uxopian AI chat panel to a FlowerDocs deployment running locally with Docker Compose. The result is a FlowerDocs instance where users can open an AI chat panel via keyboard shortcut, with document context provided by the FlowerDocs tools.

Development configuration only

This quickstart uses DevProvider (authentication disabled) and is intended to validate the scope file installation quickly. It is not suitable for production. For a production deployment — with FlowerDocsProvider, token validation, and a proper LLM check before UI integration — follow the phased Integrate with FlowerDocs guide instead.

Prerequisites​

  • A running FlowerDocs deployment with Docker Compose, with a scope already created. Follow FlowerDocs' own Docker installation guide if you don't have one yet — this quickstart doesn't repeat it. You'll need the Docker network name that stack uses (flowerdocs-net in FlowerDocs' documented compose file) and the scope id you created (e.g. DEFAULT).
  • Docker and Docker Compose.
  • At least one LLM API key (see Quickstart with Docker Compose).

Step flow​

Figure: Steps from a running FlowerDocs to a working AI chat panel.

Steps​

1. Start the uxopian-ai stack on FlowerDocs' network​

Download uxopian-ai-stack.yml and gateway-application.yaml from Quickstart with Docker Compose — same files, no FlowerDocs-specific variant needed. Two changes to that compose file before starting it:

  1. Join FlowerDocs' own Docker network instead of creating a new one — replace the networks: block at the bottom with the network FlowerDocs is already running on, marked external:

    networks:
      flowerdocs-net: # the network name your FlowerDocs compose stack uses
        external: true

    and point every service's networks: list at flowerdocs-net instead of uxopian-ai-net.

  2. Name the gateway service/container gateway-service. The scope files downloaded in the next step ship a FlowerDocs Route pointing at http://gateway-service:8085 — the gateway container must resolve under that exact name on the shared network for FlowerDocs' internal proxy to reach it:

    services:
      gateway-service: # renamed from uxopian-gateway
        container_name: gateway-service
        # ...unchanged otherwise

Then start it:

docker compose -f uxopian-ai-stack.yml up -d

2. Download the scope files​

Download

uxopian-ai_flowerdocs_scope.zip

Extract the ZIP. See Configure FlowerDocs scope files for what each file does; you don't need to edit anything for this quickstart.

3. Import the scope files with CLM​

Run the same flower-docs-clm image your FlowerDocs deployment uses, on its network, targeting your scope (replace DEFAULT, the password, and the core URL with your own):

docker run --rm --network=flowerdocs-net \
  --volume="$PWD/conf:/clm/uxopian-ai-scope/conf" \
  artifactory.arondor.cloud:5001/flower-docs-clm:{{version}} \
  update-config --template=uxopian-ai-scope --scope=DEFAULT --password=<yourPassword> \
  --ws.url=http://flower-docs-core:8081/core/rest --data.dir=/clm/

update-config only touches configuration components (Routes, Scripts) — it never touches documents, so it's safe to re-run any time you update the scope files. --ws.url must end in /core/rest (not /core/services — an older path that no longer works on the 2026 line).

4. Verify​

  1. Log in to FlowerDocs.
  2. Use the keyboard shortcut assigned to OpenChatShortcut.
  3. The Uxopian AI chat panel should appear embedded in the FlowerDocs UI.
  4. Type a question and verify a response is returned from uxopian-ai.

If the chat panel does not appear, check the uxopian-ai logs:

docker compose -f uxopian-ai-stack.yml logs uxopian-ai gateway-service

Architecture in this stack​

Figure: Chat panel traffic flow from FlowerDocs through the gateway to uxopian-ai — both stacks share one Docker network.

Last updated on