Skip to content
Draft
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
29 changes: 21 additions & 8 deletions src/handbook/engineering/ops/self-hosted-assistant.md
Original file line number Diff line number Diff line change
Expand Up @@ -8,38 +8,51 @@ FlowFuse Expert is a collection of LLM based resources provided on FlowFuse Clou

Access to these features is also available to Enterprise Licensed Self Hosted and Dedicated Customers.

Self Hosted customers are directed to contact support to request an Authentication token to enable the features.
Self Hosted customers are directed to contact support to request the necessary authentication tokens to enable the features.

The FlowFuse Expert consists of two internal components that each need to be enabled with their own token. We are working to consolidate and simplify this configuration, but this is how it needs to be done for FlowFuse 2.28.

## Process

1. Customer emails `support@flowfuse.com`
2. Support needs to verify that the customer has a current Enterprise License
3. Once confirmed open the Instance Settings for the `flow-gen` instance in the `Internal Tools` Application. Under the Security settings create a new HTTP Bearer Token using the customer name as the token name. The token will only be displayed once, so make a note of it.
4. Provide the token to the customer along with instructions on how and where to include this in the configuration.
3. Once confirmed, Engineering needs to create **two** access tokens for the customer.
1. **Assistant**: Open the Instance Settings for the `flow-gen` instance in the `Internal Tools` Application. Under the Security settings create a new HTTP Bearer Token using the customer name as the token name. The token will only be displayed once, so make a note of it - this is the **Assistant Token**.
2. **Expert**: Open the Instance Settings for the `flowfuse-expert-api` instance in the `Internal Tools` Application. Under the Security settings create a new HTTP Bearer Token using the customer name as the token name. The token will only be displayed once, so make a note of it - this is the **Expert Token**.
5. Provide the both tokens to the customer along with instructions on how and where to include this in the configuration - details below.


## Configuration

### Docker

The feature is enabled by editing the `configs.flowfuse.content` section at the top of the `docker-compse.yml` file.
Add the following after the end of the `npmRegistry` section. The `assistant` and `expert` keys should be indented 6 spaces to match.

e.g. at line 42
Insert the two tokens

```
assistant:
enabled: true
service:
url: https://flow-gen.flowfuse.cloud/v1/openai
token: <Token>
token: <Assistant Token>
expert:
enabled: true
service:
url: https://flowfuse-expert-api.flowfuse.cloud/v4/expert
token: <Expert Token>
```

The `assistant` key should be indented 6 spaces

Comment on lines -37 to +47
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this should be kept, the indentation is important

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I had moved it up to the end of line 30 alongside the instructions of where to put it.


### Kubernetes

The feature is enabled by adding the token to the values passed to the Helm chart.
The feature is enabled by adding the tokens to the values passed to the Helm chart.

- `forge.assistant.enabled` should be set to `true`
- `forge.assistant.service.url` should be set to `https://flow-gen.flowfuse.cloud/v1/openai`
- `forge.assistant.service.token` should be set to the provided Token
- `forge.assistant.service.token` should be set to the provided Assistant Token
- `forge.expert.enabled` should be set to `true`
- `forge.expert.service.url` should be set to `https://flowfuse-expert-api.flowfuse.cloud/v4/expert`
- `forge.expert.service.token` should be set to the provided Expert Token
Loading