AI Assistant
AI-powered coding help, built on Claude, inside BridgeLink WebAdmin.
Why it matters
An AI pair-programmer for your channels.
Writing and debugging channel JavaScript can be slow, and building realistic test messages by hand is tedious. AI Assistant brings Claude directly into BridgeLink WebAdmin, so engineers can explain, generate, and fix code without leaving the editor, generate synthetic test messages on demand, and diagnose failed messages in plain language. It is tuned to BridgeLink's Rhino-based JavaScript runtime and to healthcare message formats, and it runs on your own Anthropic API key, so usage stays on your account and under your control.
Capabilities
What it does:
AI Assistant chat
Open a running chat about the channel, code template, or script you have open, with context-aware suggestions and a one-click script review.
Explain, generate, and fix code
Select code for a plain-language explanation, describe what you want to generate it, or review a suggested fix as a diff before applying, all tuned to BridgeLink's Rhino ES5 runtime.
Generate sample messages
Create synthetic HL7 and other sample messages for a template using fictional data only, never real patient information.
Explain message errors
Get a plain-language diagnosis of a failed message from the message browser, then jump straight to the failing connector.
Channel descriptions
Generate or improve a channel's description from its own configuration, right on the Summary and Settings tab.
Release highlights
What's new.
Editor actions
- Explain Code: select any code and get a plain-language explanation of what it does.
- Generate Code: describe what you want in natural language and get working code inserted into the editor.
- Fix / Improve Code: review a diff of the suggested rewrite, refine it, then accept or reject.
- AI Assistant chat: ask context-aware questions about the channel, code template, or script you have open.
Message tools
- Generate Sample Messages: synthetic test messages for Message Templates, using fictional data only.
- Explain Message Errors: plain-language diagnosis of a failed message from the message browser.
- Channel descriptions: generate or improve a channel's description from its own configuration.
Privacy and control
- One org-wide API key: your Anthropic API key is entered once and shared by all users.
- Model picker: choose from Anthropic's current Claude model catalog, refreshed live.
- Privacy by default: message templates are excluded from AI prompts unless you opt in.
- Tuned for BridgeLink: aware of the Rhino (ES5) JavaScript runtime and healthcare message formats.
How it works
Get started in minutes.
- 1
License and enable the plugin
Innovar Healthcare installs the AI Assistant extension and a license. Once licensed, an AI Assistant tab appears under Settings in WebAdmin.
- 2
Add your Anthropic API key
Enter your organization's Anthropic (Claude) API key once on the AI Assistant tab, pick a Claude model, and click Test to confirm the connection. The key is shared across all users.
- 3
Work in the editor
Open any JavaScript step, code template, or global script and use Explain, Generate, or Fix/Improve from the right-click menu, or open the floating chat for context-aware help.
- 4
Generate messages and descriptions
Create synthetic sample messages from the Message Templates pane, diagnose failed messages from the message browser, and generate channel descriptions, all from within WebAdmin.
Requirements
What you'll need:
- BridgeLink 26.6.0 or later
- An AI Assistant license and the License Manager plugin
- Your organization's Anthropic (Claude) API key
- Outbound network access from BridgeLink to the Anthropic API (api.anthropic.com)
FAQ
Frequently asked questions.
Which AI models does it use?
Claude models from Anthropic. Once your API key is configured, the model list is pulled live from Anthropic's catalog, and you pick the model AI Assistant uses.
Do I need my own account?
Yes. You supply your organization's own Anthropic (Claude) API key, entered once and shared across users, so usage runs on your account and under your control.
Where does it work inside BridgeLink?
In the JavaScript editors for channels, code templates, and global scripts, and in the Message Templates pane, the message browser, and the channel Summary and Settings tab.
Is patient data sent to the AI?
Only when you choose. Message templates are excluded from prompts by default, and before diagnosing a failed message AI Assistant asks you to confirm. Generated sample messages always use fictional data.
Will generated code run on BridgeLink?
AI Assistant is tuned to BridgeLink's Rhino ES5 runtime and warns you when a suggestion uses modern JavaScript syntax, so you can adjust it before applying.
Availability
Requires BridgeLink 26.6.0 or later, an AI Assistant license, and your organization's own Anthropic (Claude) API key. Installed by Innovar Healthcare or your professional services contact.
Add AI Assistant to BridgeLink.
Talk to us about your deployment, or read the full user guide to get started.