llms.txt
# LLMTest
> LLMTest is an AI proxy for solo devs and indie hackers. Route your LLM calls through our OpenAI-compatible endpoint to get cost tracking, automatic fallbacks, JSON recovery, and autopilot prompt optimization across 340+ models.
LLMTest helps you cut LLM API costs without breaking your prompts. You write a rough prompt and ship. Every week we run shorter and cheaper variants against your real traffic. Wins that clear every safety gate go live. One click reverts anything.
## Core pages
- [Homepage](https://llmtest.io/): product overview and pricing
- [Models](https://llmtest.io/models): directory of 340+ LLMs with prices and context windows, updated daily
- [Quickstart](https://llmtest.io/docs/quickstart): get an API key and make the first call in under a minute
- [Proxy / API Gateway](https://llmtest.io/docs/proxy): OpenAI-compatible endpoint documentation
- [MCP Setup](https://llmtest.io/docs/mcp-setup): connect LLMTest to Claude Code, Cursor, Windsurf
- [Benchmarks](https://llmtest.io/docs/benchmarks): how LLMTest benchmarks and judges models
- [Autopilot](https://llmtest.io/docs/benchmarks): weekly auto-optimization with safety gates
## API
- OpenAI-compatible endpoint: `https://llmtest.io/v1/chat/completions`
- Sitemap: https://llmtest.io/sitemap.xml
- Blog RSS: https://llmtest.io/blog/rss.xml
## Blog
- [How to add per-request token budget controls in Node.js 2026](https://llmtest.io/blog/per-request-token-budget-controls-nodejs-2026): Count tokens before every LLM API call, set max_tokens per Express route, and recover cleanly from streaming truncation without surfacing errors to users.
- [Best LLMs for image analysis in 2026: $0.25 to $26 per 1K images](https://llmtest.io/blog/best-llms-image-analysis-2026): Five vision-capable LLMs compared for product photo captions, receipt parsing, and chart understanding in 2026, with per-1,000-image cost math.
- [Multi-turn chat API in Node.js 2026: history and token budgets](https://llmtest.io/blog/multi-turn-chat-api-nodejs-2026): Build a multi-turn Node.js chat API with server-side message history, context compaction, and hard per-session token budget enforcement in under 90 lines.
- [How to retry LLM API calls in Node.js in 2026: three-layer pattern](https://llmtest.io/blog/how-to-retry-llm-api-calls-nodejs-2026): Client backoff with jitter, proxy fallback, and dead-letter queue: the three Node.js layers that stop cascading LLM failures in production.
- [What is prompt injection? The 3 attack surfaces to fix](https://llmtest.io/blog/prompt-injection-explained): Prompt injection hijacks your LLM via user input, retrieved docs, and tool results. Three attack surfaces, and the defenses that actually work in production.
- [LLM cost per user in 2026: $0.15–$0.73 across three workloads](https://llmtest.io/blog/llm-cost-per-user-2026): Per-user LLM cost at 1,000 daily users in 2026: token math across chat, extraction, and RAG at two model tiers, with break-even math for solo devs.
- [How to add LLM request queuing in Node.js 2026: no Redis](https://llmtest.io/blog/llm-request-queuing-nodejs-2026): An in-process LLM request queue for Node.js: concurrency cap, per-user fairness, and backpressure under 60 lines. No Redis or BullMQ needed in 2026.
- [How to cache LLM API calls in Node.js 2026: Redis and break-even math](https://llmtest.io/blog/llm-api-response-caching-redis-nodejs-2026): Redis exact-match caching cuts LLM API spend at scale. Working Node.js middleware, 30-minute TTL trade-off, and when semantic search is worth the overhead.
- [What is an AI agent? Tool loops, stalls, and 3 prod decisions](https://llmtest.io/blog/what-is-an-ai-agent): An AI agent is an LLM in a tool-calling loop. How loops work, why they stall or run away, and the 3 architecture decisions that matter in production.
- [How to add per-user LLM cost guardrails in Node.js in 2026](https://llmtest.io/blog/per-user-llm-cost-guardrails-nodejs-2026): Add per-user LLM budget caps with Redis: soft warnings at 80%, hard stops at 100%, and automatic model downgrade. Full Node.js middleware under 60 lines.
- [LLM fine-tuning cost in 2026: $3 to $30 self-hosted, break-even math](https://llmtest.io/blog/llm-fine-tuning-cost-2026): Fine-tuning LLMs in 2026 costs $3-$30 in GPU time for open-weight models, or a 50% inference premium on GPT-4.1. Break-even math across three real scenarios.
- [How to build a RAG API in Node.js 2026: pgvector and cited answers](https://llmtest.io/blog/rag-api-nodejs-pgvector-2026): Build a production RAG API in Node.js using pgvector: text ingestion, chunked embeddings, vector search, and source-cited answers in under 80 lines.
- [LLM hallucinations explained: causes, types, production fixes](https://llmtest.io/blog/llm-hallucinations-explained): Why LLMs hallucinate, the three error types, prompting patterns that reduce them, and the one production eval that catches failures before your users see them.
- [JSON mode explained: when schema enforcement beats prompting](https://llmtest.io/blog/json-mode-explained): JSON mode vs structured output: how constrained decoding works, when 3% failure rates become a real cost, and which enforcement approach your pipeline needs.
- [How to add context management to a Node.js chat API in 2026](https://llmtest.io/blog/llm-context-management-nodejs-chat-api-2026): Add token counting, sliding window, and summarization compaction to a Node.js chat route to keep conversations coherent without blowing the context budget.
