[need help] blog update: ai agent blog #3169

LinaLam · 2025-01-25T01:12:26Z

updated with how to integrate for most the platforms

vercel · 2025-01-25T01:12:30Z

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
helicone	❌ Failed (Inspect)			Jan 28, 2025 7:18am
helicone-bifrost	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Jan 28, 2025 7:18am
helicone-eu	❌ Failed (Inspect)			Jan 28, 2025 7:18am

fumedev · 2025-01-25T01:16:06Z

Summary

Added integration guides for LLM observability with Dify, LlamaIndex, CrewAI, and LangChain
Reorganized comparison links under "Other Comparisons" sections for better structure
Added detailed code examples and configuration steps for integrating Helicone with each framework
Updated formatting and styling of framework comparisons to improve readability

🔍 Fume is reviewing this PR!

🔗 Track the review progress here:
https://app.fumedev.com/chat/e79ce617-1767-4cf9-885a-511c1a68c7ee

fumedev · 2025-01-25T01:17:08Z

bifrost/app/blog/blogs/ai-agent-builders/src.mdx

+
+2. Set up environment variables
+
+```


Please specify the language for this code block by adding bash at the start

🔨 See the suggested fix

@@ -268,11 +268,7 @@ | • **Scalability:** Useful for prototyping complex AI solutions and large datasets efficiently. | • **Poor documentation:** often outdated or unclear, however, has a large active community for support. | | • **Integrations:** Has many integrations which requires more coding effort, but offers flexibility for custom models. | • **Not suitable for production environments**: due to instability and frequent changes. | -### Integrating LLM Observability with LangChain - -1. Create an <a href="https://www.helicone.ai/signup" target="_blank" rel="noopener">Helicone</a> account, then generate a `write-only` API key. - -2. Set up environment variables +3. Modify your LangChain OpenAI configuration to use Helicone:

export HELICONE_API_KEY=your_helicone_api_key

[![Click Here](https://fume.b-cdn.net/commit_changes.svg)](https://app.fumedev.com/commit/ad75e85a-be22-4e9b-b1da-c280637284df) ⚠️ Fume is an LLM-based tool and can make mistakes. </details>

fumedev · 2025-01-25T01:18:09Z

bifrost/app/blog/blogs/ai-agent-builders/src.mdx

+
+2. Set up environment variables
+
+```bash


For consistency with other sections, please use an f-string here instead of string concatenation: f'https://oai.helicone.ai/{HELICONE_API_KEY}/v1'

🔨 See the suggested fix

@@ -268,11 +268,8 @@ | • **Scalability:** Useful for prototyping complex AI solutions and large datasets efficiently. | • **Poor documentation:** often outdated or unclear, however, has a large active community for support. | | • **Integrations:** Has many integrations which requires more coding effort, but offers flexibility for custom models. | • **Not suitable for production environments**: due to instability and frequent changes. | -### Integrating LLM Observability with LangChain - -1. Create an <a href="https://www.helicone.ai/signup" target="_blank" rel="noopener">Helicone</a> account, then generate a `write-only` API key. - -2. Set up environment variables +- All LLM calls made by the agent will be automatically logged and monitored by Helicone. For Antrhopic's Claude, a similar approach can be used. +- See <a href="https://docs.helicone.ai/integrations/openai/langchain" target="_blank" rel="noopener">OpenAI LangChain docs</a> or <a href="https://docs.helicone.ai/integrations/anthropic/langchain" target="_blank" rel="noopener">Anthropic LangChain docs</a> for details.

export HELICONE_API_KEY=your_helicone_api_key

[![Click Here](https://fume.b-cdn.net/commit_changes.svg)](https://app.fumedev.com/commit/addf876a-5058-4fb7-862a-d028ee758e61) ⚠️ Fume is an LLM-based tool and can make mistakes. </details>

fumedev · 2025-01-25T01:22:36Z

bifrost/app/blog/blogs/ai-agent-builders/src.mdx

+
+```javascript
+const llm = new OpenAI({
+  modelName: "gpt-3.5-turbo",


The OpenAI API configuration property should be 'model' instead of 'modelName'

🔨 See the suggested fix

@@ -268,11 +268,20 @@ | • **Scalability:** Useful for prototyping complex AI solutions and large datasets efficiently. | • **Poor documentation:** often outdated or unclear, however, has a large active community for support. | | • **Integrations:** Has many integrations which requires more coding effort, but offers flexibility for custom models. | • **Not suitable for production environments**: due to instability and frequent changes. | -### Integrating LLM Observability with LangChain - -1. Create an <a href="https://www.helicone.ai/signup" target="_blank" rel="noopener">Helicone</a> account, then generate a `write-only` API key. - -2. Set up environment variables +3. Modify your LangChain OpenAI configuration to use Helicone: + +const llm = new OpenAI({ + model: "gpt-3.5-turbo", + configuration: { + basePath: "https://oai.helicone.ai/v1", + defaultHeaders: { + "Helicone-Auth": `Bearer ${process.env.HELICONE_API_KEY}`, + }, + }, +}); + +- All LLM calls made by the agent will be automatically logged and monitored by Helicone. For Antrhopic's Claude, a similar approach can be used. +- See <a href="https://docs.helicone.ai/integrations/openai/langchain" target="_blank" rel="noopener">OpenAI LangChain docs</a> or <a href="https://docs.helicone.ai/integrations/anthropic/langchain" target="_blank" rel="noopener">Anthropic LangChain docs</a> for details.

export HELICONE_API_KEY=your_helicone_api_key

[![Click Here](https://fume.b-cdn.net/commit_changes.svg)](https://app.fumedev.com/commit/8a1aa6c3-6916-49d8-b514-632466fae10c) ⚠️ Fume is an LLM-based tool and can make mistakes. </details>

fumedev · 2025-01-25T01:24:10Z

bifrost/app/blog/blogs/ai-agent-builders/src.mdx

+  configuration: {
+    basePath: "https://oai.helicone.ai/v1",
+    defaultHeaders: {
+      "Helicone-Auth": `Bearer ${process.env.HELICONE_API_KEY}`,


Please use 'Helicone-Auth-Token' instead of 'Helicone-Auth' as the authentication header

🔨 See the suggested fix

@@ -215,9 +215,7 @@ ### Integrating LLM Observability with CrewAI -1. Create an <a href="https://www.helicone.ai/signup" target="_blank" rel="noopener">Helicone</a> account, then generate a `write-only` API key. - -2. Set up environment variables +3. Modify your LangChain OpenAI configuration to use Helicone: ```bash import os @@ -281,23 +279,8 @@ 3. Modify your LangChain OpenAI configuration to use Helicone: ```javascript -const llm = new OpenAI({ - modelName: "gpt-3.5-turbo", - configuration: { - basePath: "https://oai.helicone.ai/v1", - defaultHeaders: { - "Helicone-Auth": `Bearer ${process.env.HELICONE_API_KEY}`, - }, - }, -}); -``` - - All LLM calls made by the agent will be automatically logged and monitored by Helicone. For Antrhopic's Claude, a similar approach can be used. - See <a href="https://docs.helicone.ai/integrations/openai/langchain" target="_blank" rel="noopener">OpenAI LangChain docs</a> or <a href="https://docs.helicone.ai/integrations/anthropic/langchain" target="_blank" rel="noopener">Anthropic LangChain docs</a> for details. - -### Other LangChain Comparisons: - -- **<a href="https://www.helicone.ai/blog/llamaindex-vs-langchain" rel="noopener" target="_blank">LangChain vs. LlamaIndex</a>** ---

_{⚠️ Fume is an LLM-based tool and can make mistakes.}

fumedev · 2025-01-25T01:26:14Z

bifrost/app/blog/blogs/ai-agent-builders/src.mdx

@@ -87,7 +87,15 @@ Dify AI is an open-source platform that simplifies the development of AI agents
 | • **Strong data security**: Robust encryption and protection mechanisms to ensure data confidentiality and safety.               | • **Complex data processing**: Constraints in handling intricate machine learning models or extensive computational tasks.                                                                   |
 | • **Seamless integration**: Connects with popular AI models and supports integration with external tools like Zapier, Make, etc. | • **Scalability**: While there are limitations when it comes to building highly complex or large-scale tasks, Dify is suitable for building most AI apps.                                    |

-**You might be interested in**: <a href="https://www.helicone.ai/blog/crewai-vs-dify-ai" rel="noopener" target="_blank">Dify vs. CrewAI</a>.
+### Integrating LLM Observability with Dify


Please add specific steps on how to configure the API Base in Dify's UI (where to find it, what value to enter) and include a verification step to ensure the integration is working correctly

🔨 See the suggested fix

@@ -91,7 +91,19 @@ 1. Create an <a href="https://www.helicone.ai/signup" target="_blank" rel="noopener">Helicone</a> account, then generate an API key. -2. Configure API Base in Dify to use Helicone. See <a href="https://docs.helicone.ai/other-integrations/dify" target="_blank" rel="noopener">Dify Integration docs</a> for details. +2. Configure API Base in Dify to use Helicone: + - Navigate to your Dify dashboard + - Go to "Settings" > "Model Providers" + - Select "OpenAI" or add it if not present + - In the "API Base" field, enter: `https://oai.helicone.ai` + - Add your Helicone API key to the "Authorization" header with format: `Bearer YOUR_HELICONE_API_KEY` + - Click "Save" to apply the changes + +3. Verify the integration: + - Create a test conversation in your Dify application + - Send a few messages to generate LLM calls + - Visit your Helicone dashboard + - Check the "Requests" section to confirm the calls are being logged ### Other Dify Comparisons:

_{⚠️ Fume is an LLM-based tool and can make mistakes.}

fumedev · 2025-01-25T01:32:04Z

Testing

Created and ran CrewAI integration test:
- Environment variable setup works correctly
- API URL format with auth token is valid
- All tests passed
Unable to test LangChain and LlamaIndex integrations due to Python package installation restrictions.

Issues Found (through code review)

LangChain JavaScript config:
- modelName property should be model
- Helicone-Auth header should be Helicone-Auth-Token
API endpoint URLs need updating from helicone.ai to hconeapi.com
Dify integration lacks specific configuration steps, only links to docs

LinaLam added 2 commits January 23, 2025 17:48

made some changes

5ccfc9c

added integration methods

2b4f3bb

vercel bot deployed to Preview – helicone-eu January 25, 2025 01:12 View deployment

LinaLam changed the title ~~blog update: Optimize ai agents~~ blog update: ai agent blog Jan 25, 2025

vercel bot deployed to Preview – helicone January 25, 2025 01:15 View deployment

fumedev bot reviewed Jan 25, 2025

View reviewed changes

LinaLam added 2 commits January 27, 2025 23:04

Merge branch 'main' into optimize-ai-agents

2a964ce

changed the code snippet

56f08be

vercel bot deployed to Preview – helicone-bifrost January 28, 2025 07:12 View deployment

vercel bot had a problem deploying to Preview – helicone-eu January 28, 2025 07:14 Failure

vercel bot had a problem deploying to Preview – helicone January 28, 2025 07:18 Failure

LinaLam changed the title ~~blog update: ai agent blog~~ [need help] blog update: ai agent blog Jan 28, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[need help] blog update: ai agent blog #3169

[need help] blog update: ai agent blog #3169

LinaLam commented Jan 25, 2025 •

edited

Loading

vercel bot commented Jan 25, 2025 •

edited

Loading

fumedev bot commented Jan 25, 2025

fumedev bot Jan 25, 2025

fumedev bot Jan 25, 2025

fumedev bot Jan 25, 2025

fumedev bot Jan 25, 2025

fumedev bot Jan 25, 2025

fumedev bot commented Jan 25, 2025

[need help] blog update: ai agent blog #3169

Are you sure you want to change the base?

[need help] blog update: ai agent blog #3169

Conversation

LinaLam commented Jan 25, 2025 • edited Loading

vercel bot commented Jan 25, 2025 • edited Loading

fumedev bot commented Jan 25, 2025

Summary

fumedev bot Jan 25, 2025

Choose a reason for hiding this comment

fumedev bot Jan 25, 2025

Choose a reason for hiding this comment

fumedev bot Jan 25, 2025

Choose a reason for hiding this comment

fumedev bot Jan 25, 2025

Choose a reason for hiding this comment

fumedev bot Jan 25, 2025

Choose a reason for hiding this comment

fumedev bot commented Jan 25, 2025

Testing

Issues Found (through code review)

LinaLam commented Jan 25, 2025 •

edited

Loading

vercel bot commented Jan 25, 2025 •

edited

Loading