generator-ai Error: Load model (Protobuf parsing failed)

[mrfelton]

I tried this:

Hi, I'm unable to use the generator-api plugin (with google)

With google config only

When supplying a google api key, the generator fails, complaining that we don't have an openapi key, even though I'm not wanting to use it with openapi.

npm run generate

.env:

GOOGLE_GENERATIVE_AI_API_KEY=[MY GOOGLE KEY]

Error loading plugin: OpenAIError: The OPENAI_API_KEY environment variable is missing or empty; either provide it, or instantiate the OpenAI client with an apiKey option, like new OpenAI({ apiKey: 'My API Key' }).
at new OpenAI (/Users/tom/workspace/zap-middleware/eventcatalog/node_modules/openai/index.js:72:19)
at Object. (/Users/tom/workspace/zap-middleware/eventcatalog/node_modules/@eventcatalog/core/node_modules/@eventcatalog/generator-ai/dist/index.js:1207:14)

With some dummy value for openapi:

If I add some dummy value for the open api key, I get a different error:

npm run generate

.env:

OPENAI_API_KEY=dummy
GOOGLE_GENERATIVE_AI_API_KEY=[MY GOOGLE KEY]

@eventcatalog/generator-ai processing your catalog...

• Splitting markdown documents...
  Error loading plugin: Error: Load model from /Users/tom/workspace/zap-middleware/eventcatalog/node_modules/@huggingface/transformers/.cache/Xenova/all-MiniLM-L6-v2/onnx/model.onnx failed:Protobuf parsing failed.
  at new OnnxruntimeSessionHandler (/Users/tom/workspace/zap-middleware/eventcatalog/node_modules/onnxruntime-node/dist/backend.js:24:92)
  at Immediate. (/Users/tom/workspace/zap-middleware/eventcatalog/node_modules/onnxruntime-node/dist/backend.js:64:29)
  at process.processImmediate (node:internal/timers:483:21)

This happened:

I'm unable to se the chat plugin (generator-ai) with a Google API key. It requires an OpenAPI key. Using a dummy value for the open api key then results in an error.

I expected this:

I expect to be able to use the chat plugin (generator-ai) with a Google API key.

Is there a workaround?

No response

Anything else?

My config:

{
    generators: [
        [
            "@eventcatalog/generator-ai", {
                splitMarkdownFiles: true,
                includeUsersAndTeams: false,
                includeSchemas: true,
                includeCustomDocumentation: false
            },
        ],
    ],
    chat: {
        enabled: true,
        provider: 'google',
        model: 'gemini-2.5-flash-preview-04-17',
    },
    output: 'server'
}

EventCatalog Version

2.42.1

Node.js Version

v20.18.1

Platform(s)

MacOS

Community Notes

• Please vote by adding a 👍 reaction to the issue to help us prioritize.
• If you are interested to work on this issue, please leave a comment.