Documentation · Package · npm
The LLM call framework. Resilience, observability, and control for every call.
One interface across OpenAI-compatible, Anthropic, Gemini, and Bedrock, with retry budgets, circuit breaking, multi provider fallback, rate limiting, caching, and middleware built in, all running in your own process rather than a new network hop.
import Anthropic from '@anthropic-ai/sdk';
import OpenAI from 'openai';
import { VernLLM } from 'vern-llm';
import { fromAnthropic, fromOpenAI } from 'vern-llm/adapters';
const openai = fromOpenAI(new OpenAI({ apiKey: process.env.OPENAI_API_KEY }));
const anthropic = fromAnthropic(new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY }));
const llm = new VernLLM({
client: openai,
model: 'gpt-6-sol',
fallback: { client: anthropic, model: 'claude-opus-5-5', circuitBreaker: true },
rateLimit: { requestsPerMinute: 500, tokensPerMinute: 100_000, maxConcurrent: 20 },
retryBudget: { windowMs: 60_000, minCalls: 20, retryRatio: 0.2 },
maxRetries: 3,
timeoutMs: 10_000,
defaultMaxTokens: 1000,
defaultReasoningEffort: 'medium',
});
const result = await llm.call({ userContent: "What's the weather in New York?" });Works with OpenAI, Groq, Mistral, DeepSeek, Cerebras, Together AI, Fireworks AI, Ollama, Anthropic, Gemini, AWS Bedrock (via vern-llm-bedrock), or any provider reachable over HTTP via a fromFetch adapter.
This is a pnpm monorepo with five workspaces, managed with Nx for task caching and affected-based CI:
| Path | Description |
|---|---|
packages/vern-llm |
The vern-llm npm package: source, tests, and its own README with the full API reference. |
packages/vern-llm-redis |
The vern-llm-redis npm package: Redis backed circuit breaker, rate limiter, and cache adapters for vern-llm. |
packages/vern-llm-otel |
The vern-llm-otel npm package: OpenTelemetry traces and metrics for vern-llm. |
packages/vern-llm-bedrock |
The vern-llm-bedrock npm package: the AWS Bedrock Converse adapter for vern-llm, built on the AWS SDK client. |
apps/docs |
The Fumadocs-powered documentation site. |
MIT © LakBud
