Framework Integrations

Observyze provides Proxy Gateway integrations for popular frameworks like LangChain, LlamaIndex, and the Vercel AI SDK. By routing through our proxy or using our SDK wrappers, model calls and agent workflows are automatically traced.

Dedicated Provider Integration Guides

OpenAI

OpenAI & Azure OpenAI

Trace OpenAI and Azure OpenAI endpoints with token and latency tracking.

View integration

Anthropic

Anthropic Claude

Capture Claude 3.5 Sonnet & Haiku messages, tool calls, and spend.

View integration

Google AI

Google Gemini

Trace Gemini Flash and Pro models with multimodal attribution.

View integration

Framework

LangGraph Multi-Agent

Trace complex stateful multi-agent graphs and cycle executions.

View integration

Local AI

Ollama Local Models

Monitor on-premise Llama, Mistral, and DeepSeek inference locally.

View integration

Automation

CI/CD & GitHub Actions

Automate regression testing in CI workflows using the API.

View integration

LangChain Integration

The easiest way to trace LangChain is to point the base URL to the Observyze proxy.

TypeScript Example

import { ChatOpenAI } from "@langchain/openai";

const model = new ChatOpenAI({
  configuration: {
    baseURL: "https://api.observyze.com/api/v1/proxy/openai/v1",
    defaultHeaders: {
      "Authorization": "Bearer ob_your_api_key",
      "x-project-id": "proj_123"
    }
  }
});

Vercel AI SDK

Use the custom provider setup to route Vercel AI SDK traffic through Observyze.

Next.js Route Handler

import { createOpenAI } from '@ai-sdk/openai';
import { generateText } from 'ai';

const observyzeProvider = createOpenAI({
  baseURL: 'https://api.observyze.com/api/v1/proxy/openai/v1',
  headers: {
    Authorization: `Bearer ${process.env.OBSERVYZE_API_KEY}`,
    'x-project-id': process.env.OBSERVYZE_PROJECT_ID
  }
});

export async function POST(req: Request) {
  const { text } = await generateText({
    model: observyzeProvider('gpt-4o'),
    prompt: 'Write a poem about tracing',
  });
  
  return Response.json({ text });
}

Ollama (Local Models)

Trace local open-source models (Llama 3.2, DeepSeek-R1, Mistral, Qwen 2.5, Phi-3) running on your local machine with zero API cost attribution.

TypeScript SDK (OpenAI-compatible)

import OpenAI from "openai";
import { ObservyzeClient } from "@observyze/sdk";

const observyze = new ObservyzeClient({
  apiKey: process.env.OBSERVYZE_API_KEY!,
  projectId: process.env.OBSERVYZE_PROJECT_ID!,
});

// Point OpenAI client at local Ollama instance
const openai = observyze.wrapOpenAI(
  new OpenAI({
    baseURL: "http://localhost:11434/v1",
    apiKey: "ollama", // Required by SDK, ignored by Ollama
  })
);

const response = await openai.chat.completions.create({
  model: "llama3.2",
  messages: [{ role: "user", content: "Summarize this agent task" }],
});

console.log(response.choices[0].message.content);

Python (LangChain / OpenAI-compatible)

from openai import OpenAI
from observyze import ObservyzeClient

observyze = ObservyzeClient(
    api_key="ob_your_api_key",
    project_id="proj_123"
)

# Instrument local Ollama
client = observyze.wrap_openai(
    OpenAI(base_url="http://localhost:11434/v1", api_key="ollama")
)

response = client.chat.completions.create(
    model="llama3.2",
    messages=[{"role": "user", "content": "Explain circuit breakers in distributed systems"}]
)

print(response.choices[0].message.content)