| Product | Hopscotch AI |
| Website | hopscotchlabs.ai |
| Category | LLM Router / AI Gateway |
| Pricing | Prepaid balance, pay per token at provider rates |
| Compatibility | OpenAI SDK compatible (change base URL only) |
| Models | 500+ from Anthropic, OpenAI, Google, Meta, and more |
What Is Hopscotch AI?
Hopscotch AI is an LLM router that sits between your application and every major model provider. You point your existing OpenAI SDK at their API endpoint, swap in a Hopscotch API key, and suddenly you can call Claude, GPT, Gemini, Llama, and hundreds of other models without changing anything else in your code.
The value proposition is consolidation. Instead of managing separate API keys, billing accounts, and rate limits for Anthropic, OpenAI, Google, and whoever else you need, you manage one. One key, one prepaid balance, one dashboard.
Key Features
OpenAI-Compatible Gateway
The migration path is deliberately simple. If your application already uses the OpenAI SDK, you change two things: the base URL to api.hopscotchlabs.ai and the API key. Everything else — your prompt structure, streaming logic, function calling — stays the same. Then you change the model name in your requests to access any model in their catalog.
500+ Model Catalog
Hopscotch aggregates models from all major providers: Anthropic (Claude family), OpenAI (GPT family), Google (Gemini), Meta (Llama), and a long tail of specialized and open-source models. Their /models endpoint lists everything available, and the catalog syncs hourly with provider updates.
Spend Controls
You can set spend limits per API key, per team member, and per workspace. This is genuinely useful for teams where multiple developers or projects hit the same account. You top up a prepaid balance and allocate budgets without worrying about surprise bills from a runaway prompt loop.
Chat Completions and Responses API
Hopscotch supports both the Chat Completions API (the standard for most applications) and the newer Responses API. This means newer SDK features work through the gateway without waiting for Hopscotch to implement proprietary wrappers.
Pricing
Hopscotch uses a prepaid balance model. You add credit to your account, and each API call deducts from that balance at the provider’s published rate (or close to it — Hopscotch takes a margin). There are no monthly minimums, no per-seat charges, and no tiered feature gates.
The practical benefit: if you’re a small team spending $200/month across three providers, you consolidate that into one balance with one invoice instead of three. If you’re a larger team, the spend controls and workspace features justify the routing margin.
Who Is This For?
- Teams evaluating multiple models — switch models by changing a string, not by rewriting your integration
- Startups that don’t want vendor lock-in — your code stays provider-agnostic
- Companies with budget governance requirements — per-key and per-workspace spend limits
- Developers building multi-model applications — use Claude for reasoning, GPT for coding, Gemini for vision, all through one client
Pros
- Drop-in replacement for OpenAI SDK (two-line change)
- 500+ models from all major providers
- Granular spend controls per key/team/workspace
- No monthly minimums or seat-based pricing
- Hourly catalog sync keeps models current
- Single invoice instead of multiple provider bills
Cons
- Adds a routing hop (latency cost, however small)
- Margin on top of provider rates increases per-token cost
- Provider-specific features may not be fully supported through the gateway
- Prepaid model means estimating spend upfront
- Dependency on a third-party gateway for production traffic
Verdict
Hopscotch AI solves a real operational problem: managing multiple LLM providers is genuinely tedious. Separate accounts, separate billing, separate rate limits, separate SDKs for provider-specific features. Hopscotch collapses all of that into one integration point.
The trade-off is straightforward: you pay a margin on top of provider rates for the convenience of consolidation. For teams spending $50–$500/month across two or three providers, that margin is worth the simplified billing and the ability to switch models without code changes. For teams with heavy single-provider usage, going direct will always be cheaper.
The OpenAI compatibility is the key decision: if your stack already uses the OpenAI SDK, the migration is trivially easy. If you’ve built around provider-specific features (Anthropic’s tool use, Google’s grounding), check whether Hopscotch fully supports those features before committing.
Build a developer tool?
Get reviewed and linked from our 27-site network. Placements live in 48 hours.
See Packages