# RouterPlus (slim)

> RouterPlus is an LLM gateway that puts Claude, GPT and decision models behind one API key and a prepaid balance, with OpenAI-compatible and Anthropic-compatible endpoints at the providers' list prices.

- Full: https://www.anchorterminal.com/tools/routerplus.md (~7,550 tokens) · this version ~1,830 tokens · JSON https://www.anchorterminal.com/tools/routerplus.json · canonical https://www.anchorterminal.com/tools/routerplus
- Index: https://www.anchorterminal.com/llms.txt · API: https://www.anchorterminal.com/api/v1/index.json · Updated: 2026-10-14

**D · 51.2/100 · rank #777 of 1010 · #17 in Model APIs & inference · not agent-ready · confidence medium**


Assessment: The docs are written for agents, every model price is public, and each response carries its own cost. No terms of service, privacy policy, status page or changelog were found. Signup adds no free credit, and the chat catalogue is Claude and GPT only.

## Facts

- Kind: Model router · vendor: RouterPlus · category: Model APIs & inference · legal entity: EVO HQ Inc · provenance 50/100
- Endpoint: `https://api.routerplus.com/v1` (HTTP)
- Auth: API key · pricing: Pay per use · x402: no · licence: Closed service. No terms of service for the gateway were found on routerplus.com or app.routerplus.com on 13 October 2026. The only terms on the site govern the capacity request form and name EVO HQ Inc, a Delaware corporation (https://routerplus.com/capacity/terms/).
- Probe metrics: not measured yet (probes haven't run)
- Surface graded: The gateway at https://api.routerplus.com, read from its docs (llms-full.txt), public catalogue and site on 13 October 2026. No account was created and no call was made. An unauthenticated `GET /v1/models` answered 401
- Endpoints: `POST /v1/chat/completions`, `POST /v1/messages` with `count_tokens`, `POST /v1/responses` (OpenAI models, other models in preview), `POST /v1/images/generations`, `POST /v1/decisions`, `GET /v1/models`, `GET /v1/usage`, `GET /v1/generation`, `POST /v1/route` (explains a route without dispatching) and `GET /v1/limits`
- Not available: Embeddings, audio, moderation, files and batches return 404. No idempotency keys, no model fallback lists, no sorting providers by price, and `n` above 1 is a 400
- Rate limits: 20 requests a minute for the first key and for every key until the organisation's first purchase, then 300 a minute per console key. 1,000,000 tokens a minute and 8 requests in flight per organisation and per key by default. No self-serve raise. Request bodies over 10 MB are a 413
- Errors: Every error carries a stable `error_type` in the body and in `x-tm-error-code`, with `x-tm-error-origin` saying whether a limit, a provider or the gateway refused it. 429s for rate limits carry `retry-after`; balance and spend-cap 429s (`insufficient_quota`) are not retryable
- Keys and identity: Keys are stored as SHA-256 hashes and shown once. Each key belongs to an organisation, a workspace and a principal, with per-key monthly spend caps and RPM limits that can only be lowered. Inference keys have no management rights. Owner, admin and viewer roles apply across the organisation. No SSO or SCIM
- Bring your own key: OpenAI, Anthropic and Azure keys, or an AWS Bedrock workload role, stored envelope-encrypted. The provider bills the caller directly and RouterPlus charges nothing
- Data handling: The gateway stores no prompts or completions, only token counts, timings, outcomes and cost. Two exceptions are stated. Optimize keeps eval cases and answers until the search is deleted, and accounts in a hackathon or similar programme have full requests and responses kept for up to 30 days. Upstream, Anthropic and OpenAI chat routes are marked `prompt_logging: retained` and none of them has ZDR
- Company: The site footer reads © 2026 RouterPlus and says it is backed by Y Combinator. The capacity form's terms name EVO HQ Inc, a Delaware corporation, and the Privacy and Security links are mail links to founders@evo-hq.com. EVO HQ also publishes evo, an Apache-2.0 optimisation tool, at https://github.com/evo-hq/evo
- Prices: Markup on provider token prices 0% percentage fee; Claude Opus 5.5 input $4 per 1M tokens; Claude Opus 5.5 output $20 per 1M tokens; GPT-6.1 Sol input $2 per 1M tokens; GPT-6.1 Sol output $10 per 1M tokens; Anthropic web search server tool $10 per 1,000 requests

- Scores: Reliability 45, Performance pending, Schema & documentation 59, Agent ergonomics 71, Security & auth 62, Payments & pricing 40, Task success pending, Maintenance & community 44, Transparency & trust 38 · negative events -2 · total over the 7 assessed categories
- Why: Reliability, Hosted reading. · Schema & documentation, Router, scored on the API lines. · Agent ergonomics, Router, read on the model lines (tool use, structured output, caching, context, batch, SDKs, errors). · Security & auth, Router, read on the model lines. · Payments & pricing, No machine payment protocol found in llms-full.txt. · Maintenance & community, Router, read on the model lines. · Transparency & trust, Closed service with no terms of service found.
- Sources: 17, open questions: 6, both in the full twin
- Capabilities: inference.router, inference.llm, inference.decision, image.generate
- JSON: https://www.anchorterminal.com/api/v1/tools/routerplus.json
- Verify (for the vendor): the badge `https://www.anchorterminal.com/badges/routerplus.svg` or a link to https://www.anchorterminal.com/tools/routerplus from a page on routerplus.com or one of its subdomains, then `POST https://www.anchorterminal.com/api/v1/verify` `{"slug", "url"}` or `verify_listing` at /mcp; re-checked weekly, no effect on the grade. Snippets in the full twin.

## Before you call it

1. Send `max_tokens` on every request. Without it the balance hold assumes up to 32,768 output tokens and can return `insufficient_quota` on a small balance
2. Switch on `error_type`. Retry `rate_limit` after `retry-after`, and stop on `insufficient_quota`, which needs a person to add credit or raise a cap
3. Don't add a retry loop on top of the SDK's. There is no idempotency key, so a retry that reaches a provider is billed again
4. Copy model ids from `GET /v1/models` or /api/models.json. An unknown id is a 404 and the gateway never substitutes one
5. Use `/v1/messages` for Claude when you need `thinking` or server tools, and `/v1/chat/completions` with `response_format` only for GPT ids

## Connect

```bash
curl -N https://api.routerplus.com/v1/chat/completions \
  -H "Authorization: Bearer $TM_API_KEY" \
  -H "content-type: application/json" \
  -d '{"model":"claude-haiku-4-5","stream":true,"max_tokens":60,"messages":[{"role":"user","content":"hello"}]}'
```

## Similar tools

| Tool | Grade | Score | Shared capabilities | Slim |
| --- | --- | --- | --- | --- |
| OpenRouter | B | 68.5 | inference.llm, inference.router | https://www.anchorterminal.com/tools/openrouter.min.md |
| Cloudflare Workers AI | B | 68.3 | inference.llm, image.generate | https://www.anchorterminal.com/tools/cloudflare-workers-ai.min.md |
| Cloudflare AI Gateway | B | 66.8 | inference.llm, inference.router | https://www.anchorterminal.com/tools/cloudflare-ai-gateway.min.md |
| DeepInfra | B | 63 | inference.llm, image.generate | https://www.anchorterminal.com/tools/deepinfra.min.md |
| Docker Model Runner | C | 57.1 | inference.llm, image.generate | https://www.anchorterminal.com/tools/docker-model-runner.min.md |


## Panel reviews (0, desk reviews from public material, no calls made)