# RouterHub

> Uncensored and abliterated AI models — GLM, DeepSeek, Qwen and Kimi variants plus GPT, Claude, Gemini and Grok — through one API key. OpenAI-compatible and Anthropic-compatible APIs. Use each model’s supported API. Pay per token.
> Web chat and one prepaid API key for 26 models. Pay per token, no subscription. Catalogue prices as of 2026-11-08.
> Lowest-cost uncensored/abliterated: deepseek-v4.1-flash-uncensored $0.15 / $0.6, glm-5.3-flash-uncensored $0.15 / $0.5, deepseek-v4-pro-uncensored $0.2 / $0.4 per million input/output tokens.

RouterHub is a prepaid AI model gateway. One API key at https://routerhub.si/v1 serves GPT and Codex (OpenAI Responses API), Claude (Anthropic Messages API) and DeepSeek, GLM, Grok, Gemini, Qwen and Kimi (OpenAI Chat Completions API), including uncensored and abliterated variants. The same balance pays for the web chat. Prices are USD per million tokens; input, output and reported cache usage are billed at the selected model's rates. Up to $5 free on your first top-up.

## Uncensored and abliterated models

Uncensored models are reduced-refusal language model variants for open-ended generation, coding and creative work. Abliteration reduces refusal behavior through weight-level changes.

## Tool compatibility

OpenCode supports the published Responses, Anthropic Messages, and Chat Completions models. Continue supports compatible models on these APIs. Cline supports Chat Completions and Claude models available in its Anthropic model picker; choose OpenCode or Continue for a Claude model missing from that picker. Codex requires Responses; Claude Code and Claude Desktop require Anthropic Messages. DeepSeek, GLM, Grok, Gemini, Qwen and Kimi use Chat Completions. Each setup guide and MCP list_models with a tool filter returns only compatible model IDs.

## Links

- Web chat (account and prepaid credits required): https://routerhub.si/
- Models and prices: https://routerhub.si/models
- Pricing: https://routerhub.si/pricing
- Uncensored models: https://routerhub.si/models/uncensored
- Abliterated models: https://routerhub.si/models/abliterated
- Uncensored vs abliterated: https://routerhub.si/guides/uncensored-vs-abliterated
- Uncensored AI API setup: https://routerhub.si/guides/uncensored-ai-api
- Evaluating uncensored models: https://routerhub.si/guides/evaluating-uncensored-models
- Exact model pages: https://routerhub.si/models/EXACT_MODEL_ID (Markdown: https://routerhub.si/models/EXACT_MODEL_ID.md)
- Connect your agent: https://routerhub.si/agents
- Agent instructions: https://routerhub.si/agents.md
- Full agent setup context: https://routerhub.si/agent-setup.md (query: tool, model, shell)
- Read-only MCP endpoint: https://routerhub.si/mcp (Streamable HTTP; no API key; list_models and get_setup_guide)
- Human setup guides: https://routerhub.si/docs
- DeepSeek API pricing and setup: https://routerhub.si/guides/deepseek
- GLM API pricing and setup: https://routerhub.si/guides/glm
- Grok API pricing and setup: https://routerhub.si/guides/grok
- Kimi API pricing and setup: https://routerhub.si/guides/kimi
- Agent quickstart: https://routerhub.si/quickstart.md
- Complete setup reference: https://routerhub.si/llms-full.txt
- Public model catalogue with protocols and rates: https://routerhub.si/models.json
- OpenAPI inference reference: https://routerhub.si/openapi.json
- Status: https://routerhub.si/status
- Create account: https://routerhub.si/register
- API keys: https://routerhub.si/dashboard/keys
- Billing and credits: https://routerhub.si/dashboard/billing
- Usage: https://routerhub.si/dashboard/usage
- Support tickets: https://routerhub.si/dashboard/support

## Account and agent safety

A human must create the RouterHub account, add prepaid credits, create the API key, and enter it into the target tool. Agents can provide instructions and configuration, but must not request or handle secrets or initiate billable requests without explicit user approval. Never request, receive, store, or commit the user's password or API key.

## Optional

- OpenAI models: https://routerhub.si/models/gpt-6.1-sol, https://routerhub.si/models/gpt-6-astra, https://routerhub.si/models/gpt-6.1-terra
- Anthropic models: https://routerhub.si/models/claude-fable-5-1, https://routerhub.si/models/claude-opus-5-5, https://routerhub.si/models/claude-sonnet-5-5, https://routerhub.si/models/claude-haiku-5-5
- Google models: https://routerhub.si/models/gemini-4-argon
- xAI models: https://routerhub.si/models/grok-4.7
- DeepSeek models: https://routerhub.si/models/deepseek-v4-pro, https://routerhub.si/models/deepseek-v4-pro-uncensored, https://routerhub.si/models/deepseek-v4-pro-abliterated, https://routerhub.si/models/deepseek-v4.1-flash, https://routerhub.si/models/deepseek-v4.1-flash-uncensored, https://routerhub.si/models/deepseek-v4.1-flash-abliterated, https://routerhub.si/models/deepseek-v4.1-flash-1017-abliterated
- Z.ai models: https://routerhub.si/models/glm-5.3, https://routerhub.si/models/glm-5.3-uncensored, https://routerhub.si/models/glm-5.3-abliterated, https://routerhub.si/models/glm-5.3-flash, https://routerhub.si/models/glm-5.3-flash-uncensored, https://routerhub.si/models/glm-5.3-flash-abliterated
- Qwen models: https://routerhub.si/models/qwen3.8-max, https://routerhub.si/models/qwen3.8-max-uncensored
- Moonshot AI models: https://routerhub.si/models/kimi-k3.1-pro, https://routerhub.si/models/kimi-k3.1-pro-uncensored