Helicone takes the simplest possible approach to LLM monitoring: it's a proxy. Change your OpenAI base URL from api.openai.com to oai.helicone.ai, add your Helicone API key as a header, and every LLM request is logged — latency, tokens, cost, prompts, and completions. No SDK integration, no code changes beyond a URL swap. - Source: dev.to / about 1 month ago

You're monitoring production traffic. You need Langfuse / Phoenix / Helicone / Braintrust for that. Online eval is a different problem class: implicit feedback, drift detection, hallucination rates on your data, not on HellaSwag. - Source: dev.to / about 2 months ago

For many teams, especially those starting out or with simpler needs, commercial solutions like Portkey, Helicone, OpenPipe, or LiteLLM Proxy offer off-the-shelf capabilities that cover many common proxy use cases (caching, logging, cost tracking). NeuroLink itself can be seen as an SDK that complements these, allowing you to integrate with them or build similar features on top. - Source: dev.to / 4 months ago

TL;DR: Go with Langfuse if you want open-source and self-hosted. Pick Helicone if you want the fastest setup (2 minutes, no SDK). Stick with LangSmith if your stack already runs on LangChain. And if your org already pays for Datadog, their LLM module slots right in. - Source: dev.to / 5 months ago

Hey HN, we're Justin and Cole, the founders of Helicone (https://helicone.ai) or self-deploy with our new fully open-source helm chart (https://helicone.ai/selfhost). Yet even with detailed traces, probabilistic systems are notoriously hard to debug at scale. So, we released evaluators (either via LLM-as-judge or custom Python evaluators leveraging the CodeSandbox SDK - https://codesandbox.io/docs/sdk/sandboxes).... - Source: Hacker News / over 1 year ago

Thanks for the note. Generally best to just describe the task (need to improve the system prompt to always only return tools). Here's the response I got: https://imgur.com/a/NyHBCe2 (https://programming-helper.com/ , https://explain.dev/ , https://tldrdev.ai/ , https://code-mentor.ai/) In addition to the categorization and summary (driven by GPT-4), it takes into account performance metrics of the tool (visits,... Source: about 3 years ago

Agree with so many of the comments here. I believe the way to equip folks to be productive with legacy code is build tools that replicate the goodness of an experienced engineer while on the job. Supplement the help available and ensure the person onboarding is benefitting from the questions that were asked by new folks before them. I started building the tool here: explain.dev While courses could help you feel... Source: over 3 years ago

The technology behind the images is ExplainDev, an AI powered programmer's assistant. You can think of it as an expert that's always available to answer your technical questions and explain code. - Source: dev.to / almost 4 years ago

I used explain.dev for code explanations and snappify.io for the visuals :). Source: about 4 years ago