openai-node

All main OpenAI SDK libraries are made with stainless[0][1]. [0]: https://github.com/openai/openai-python [1]: https://github.com/openai/openai-node

All main OpenAI SDK libraries are made with stainless[0][1]. [0]: https://github.com/openai/openai-python [1]: https://github.com/openai/openai-node

Basically what's described here: * https://github.com/openai/openai-node?tab=readme-ov-file#automated-function-calls * https://jrmyphlmn.com/posts/sequential-function-calls Suppose I have a model that does not natively support this. Has anyone written/came across a library that implements such logic? Conceptually, I would imagine that it simply boils down to another LLM interpreting user's question. The prompt for that LLM includes instructions that explain access to tools it has. That LLM then responds with a plan of execution. The plan is some JSON with instructions of what to execute. After the first cycle, the plan is re-evaluated. This repeats until LLM is happy with the response. In practice, this feels like _a lot_ of edge cases to handle, so I am wondering if there are existing abstractions that I could use here. For context, I am

This is the code that does the work: https://github.com/rectanglehq/Shapeshift/blob/d954dab2a866c... There are a few ways this could be made a less expensive to run: 1. Cache those embeddings somewhere. You're only embedding simple strings like "name" and "address" - no need to do that work more than once in an entire lifetime of running the tool. 2. As suggested here https://news.ycombinator.com/item?id=40973028 change the design of the tool so instead of doing the work it returns a reusable data structure mapping input keys to output keys, so you only have to run it once and can then use that generated data structure to apply the transformations on large amounts of data in the future. 3. Since so many of the keys are going to have predictable names ("name", "address" etc) you could even pre-calculate embeddings for the 1,000 most common

We try to keep it to a minimum, especially in JS (though we have some nice improvements coming soon when we deprecate node-fetch in favor of built-in fetch). The package sizes aren't tiny because we include thorough types and sourcemaps, but the bundle sizes are fairly tidy. Here's an example of a typical RESTful endpoint (Lithic's `client.cards.create()`: https://github.com/lithic-com/lithic-node/blob/36d4a6a70597e... Here are some example repos produced by Stainless: 1. https://github.com/openai/openai-node 2. https://github.com/openai/openai-python 3. https://github.com/cloudflare/cloudflare-go 4. https://github.com/Modern-Treasury/modern-treasury-java

Have you seen/tried the `.runTools()` helper? Docs: https://github.com/openai/openai-node?tab=readme-ov-file#aut... Example: https://github.com/openai/openai-node/blob/bb4bce30ff1bfb06d... (if what you're fundamentally trying to do is really just get JSON out, then I can see how json_mode is still easier).

We have a bit of context about this in the readme: https://github.com/lastmile-ai/aiconfig#what-problem-it-solv.... The main issue with keeping it in code is that it tangles application code with prompts and model-specific logic. That makes it hard to evaluate the genAI parts of the application, and also iterating on the prompts is not as straightforward as opening up a playground. Having the config be the source of truth let's you connect it to your application code (and still source controlled), lets you evaluate the config as the AI artifact, and also lets you open the config in a playground to edit and iterate. For example, compare how much simpler openai function calling becomes with storing the stuff as a config: https://github.com/lastmile-ai/aiconfig/blob/main/cookbooks/... vs using vanilla openai directly (https://github.com/open

A package I help maintain was released about a month ago, and we've had bug reports related to this: https://github.com/openai/openai-node/issues/304#issuecommen... We simply chose to mark Jest 27 as unsupported in this case, but it did cost some debugging time.

Some part of the OpenAI API can stream events, like the “chat completion” endpoint that’s used by ChatGPT to type out its response. However, the official OpenAI Node SDK doesn’t natively support streaming [1] (though it’s the most requested feature [2]). The recommended approach [3] is kind of a nightmare. I used Fern [4] to generate an SDK for the OpenAI API. It makes it a lot easier to stream responses! https://github.com/fern-openai/openai-node#streaming There’s some other niceties of the Fern-generated SDK: - Endpoints organized by resource, e.g. client.audio.transcribe(...) - Properties are camelCase to match JS/TS convention (they are snake_case in OpenAI’s SDK) - The file upload endpoints support Node.js ReadStream, rather than just File [1] https://github.com/openai/openai-node [2] https://github.com/openai/openai-node/issues?q=is

I had a look at the code. Despite the important note on https://github.com/openai/openai-node it appears to be in use here, leaking the author's API key.