Hopscotch
Hopscotch is an LLM gateway that provides a single API for accessing 500+ AI models from providers like OpenAI, Anthropic, and Google. The platform offers built-in failover routing, spend controls, and usage analytics, with zero markup on token prices and optional bring-your-own-key support. Every account includes Zero Data Retention by default, ensuring prompts and responses are never stored or used for training.
- Artificial Intelligence
- Developer Tools
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Developers building AI-powered applications face significant operational complexity when managing multiple model providers. They must handle provider outages, model failures, and rate limits manually, while also tracking costs across different APIs and ensuring consistent request routing. This fragmentation slows development and creates reliability risks for production workloads.
Solution
Hopscotch provides a unified LLM gateway that lets developers access 500+ models through a single API. The platform supports routing profiles where developers specify primary and backup models, with automatic failover to alternate providers when a model fails or experiences an outage. Requests are forwarded using the customer's own API keys when provided, and the service charges each provider's list price per token with no markup. The platform includes a comprehensive activity log that records which model answered each request, along with token counts, cost, speed, and result, accessible through both the dashboard and API.
Target Audience
Primary users are developers and engineering teams building AI applications, including AI startups, agencies, and enterprise teams that need reliable multi-provider model access with cost controls and failover capabilities.
Features
- Single API gateway for 500+ models from providers including OpenAI, Anthropic, Google, Meta, DeepSeek, and Cohere
- Routing profiles with ordered backup models and automatic failover on provider outages or request failures
- Zero Data Retention by default, with prompts and responses never stored or used for model training
- Bring-your-own-key support that sends requests on the customer's provider key with no Hopscotch fees
- Spend controls with budgets and limits for every API key and teammate, plus optional auto top-up
- Request logs showing model, provider, tokens, cost, speed, and result for every request
- Guardrails to block risky requests before they reach a model
- Usage analytics tracking spend, speed, and failure rates across all requests