---
title: "Guides"
canonical: https://uncensoredllm.ai/guides
generated: 2026-11-22T23:10:49.448Z
source: https://uncensoredllm.ai/guides
base_url: https://api.uncensoredllm.ai/v1
---

# Guides

- [Use UncensoredLLM with SillyTavern](https://uncensoredllm.ai/guides/sillytavern) (steps checked 2026-11-15): The Custom (OpenAI-compatible) source, the base URL and key, the best models for SillyTavern at live prices, settings that work and a preset to import.
- [Chub AI API key setup: proxy URL, model id, settings](https://uncensoredllm.ai/guides/chub-ai) (steps checked 2026-11-06): Put an uncensored model behind Chub AI: the reverse proxy URL and key in Secrets, the model id, the generation settings we start from, and what 100 messages cost.
- [HammerAI proxy setup with an uncensored model](https://uncensoredllm.ai/guides/hammerai-proxy) (steps checked 2026-10-24): HammerAI proxy setup in four fields with an uncensored roleplay build: the Proxy URL, the model id, the context length and the key. No GPU needed.
- [Novelcrafter custom API for uncensored fiction](https://uncensoredllm.ai/guides/novelcrafter-custom-api) (steps checked 2026-10-24): Novelcrafter custom API setup for uncensored fiction: an OpenAI Compatible Connection, prose builds with 262K+ token windows and the cost per chapter.
- [Xoul AI custom engine (proxy) with uncensored models](https://uncensoredllm.ai/guides/xoul-ai-proxy) (steps checked 2026-10-24): Xoul AI proxy setup on the Generic provider with an uncensored roleplay build, inside Xoul's 875-token and 3,500-character reply limits.
- [Agnai proxy setup: Third-Party API, uncensored model](https://uncensoredllm.ai/guides/agnai-proxy) (steps checked 2026-10-24): Agnai proxy setup with the exact labels from its source: Third-Party, OpenAI Compat (Chat), the URL, the key and the model id of an uncensored build.
- [KoboldAI Lite custom API: uncensored models, no GPU](https://uncensoredllm.ai/guides/koboldai-lite-custom-api) (steps checked 2026-10-24): KoboldAI Lite custom API setup in the browser: the OpenAI Compatible API provider, chat-completions mode, CORS proxy off and an uncensored model id.
- [KoboldAI Classic API: what it is, and the OpenAI-format route](https://uncensoredllm.ai/guides/koboldai-classic-api) (steps checked 2026-10-27): The KoboldAI Classic API is the /api/v1/generate shape KoboldCpp and SillyTavern speak. What it is, how it differs from OpenAI format, and how to put the same front-ends on an uncensored model with no GPU.
- [KoboldCpp remote API: no model loaded, uncensored builds](https://uncensoredllm.ai/guides/koboldcpp) (steps checked 2026-10-28): KoboldCpp remote API setup: launch with --nomodel, point the bundled Kobold Lite at an OpenAI-compatible URL and key, and pick an uncensored model id.
- [JannyAI: download a card, run it on any uncensored model](https://uncensoredllm.ai/guides/jannyai) (steps checked 2026-10-27): JannyAI shows and downloads Janitor AI bot definitions; it has no chat, no proxy and no API. This guide takes a JannyAI card into SillyTavern, Janitor AI or our roleplay chat on an uncensored model.
- [RisuAI custom API setup with an uncensored model](https://uncensoredllm.ai/guides/risuai-custom-api) (steps checked 2026-10-24): RisuAI custom API setup for an uncensored roleplay build: the URL, Key/Password, Request Model and Format fields, plus what a session costs.
- [Wyvern Chat API setup: provider, connection, model id](https://uncensoredllm.ai/guides/wyvern-chat-api) (steps checked 2026-11-06): Put an uncensored model behind WyvernChat on your own key: the Provider (API URL Type OpenAI, the URL, the stored key), the Connection with the model id, and what 100 messages cost.
- [Sophia's LoreBary proxy with an uncensored model](https://uncensoredllm.ai/guides/lorebary-proxy) (steps checked 2026-11-06): What LoreBary's Proxy Portal passes on (a proxy URL, your key, a model id), which of its features survive a custom proxy, and the direct route when you only want the model.
- [Add uncensored models to Open WebUI](https://uncensoredllm.ai/guides/open-webui) (steps checked 2026-10-25): Open WebUI add model steps for an uncensored catalog: one connection, a short picker set by your key, a preset per build and each build's context window.
- [LibreChat with an uncensored model](https://uncensoredllm.ai/guides/librechat) (steps checked 2026-11-06): One custom endpoint in librechat.yaml: the name, the key from .env, the base URL with /v1, a models.default list of uncensored builds and a small titleModel. The mount, the cost of a chat and the errors.
- [Ollama uncensored models: every library tag, and the same builds by API](https://uncensoredllm.ai/guides/ollama-uncensored-models) (steps checked 2026-10-27): Every ollama uncensored model on the Ollama library, read today, next to the same or newer abliterated builds on one prepaid key: size class, live price per million tokens, cost per 100 messages, and the two-line config for Open WebUI or AnythingLLM.
- [LiteLLM with an uncensored OpenAI-compatible endpoint](https://uncensoredllm.ai/guides/litellm) (steps checked 2026-10-25): LiteLLM OpenAI-compatible endpoint setup for uncensored builds: openai/ plus the model id and api_base, in the SDK and in config.yaml, with a fallback chain.
- [n8n with an uncensored model over the OpenAI-compatible API](https://uncensoredllm.ai/guides/n8n) (steps checked 2026-11-22): n8n OpenAI-compatible API setup for an uncensored model: an OpenAI credential with a custom Base URL, the OpenAI Chat Model under an AI Agent, the id typed in.
- [Build an uncensored Discord bot in one Python file](https://uncensoredllm.ai/guides/discord-bot) (steps checked 2026-10-25): An uncensored Discord bot you run yourself: one discord.py file, the OpenAI SDK on this API, four env vars, and replies only in channels you list.
- [The OpenAI SDK with uncensored models](https://uncensoredllm.ai/guides/sdks) (steps checked 2026-10-26): Set OPENAI_BASE_URL and OPENAI_API_KEY and keep your code: Python, TypeScript, curl and the Vercel AI SDK, streaming, JSON output, the error codes, and what a request costs at the live price.
- [The Vercel AI SDK with an uncensored model](https://uncensoredllm.ai/guides/ai-sdk) (steps checked 2026-11-02): createOpenAICompatible with our base URL, then generateText, streamText and tool() on an uncensored build: which builds take tool calls, the usage fields, a Next.js route handler, the error codes, and what one call costs at the live price.
- [Claude Code with an uncensored model](https://uncensoredllm.ai/guides/claude-code) (steps checked 2026-10-26): Run an uncensored build inside Claude Code with no proxy: ANTHROPIC_BASE_URL on the Anthropic-compatible base URL, the key as ANTHROPIC_AUTH_TOKEN, two model ids priced from the live catalog. claude-code-router as the optional path.
- [Cursor with an uncensored model](https://uncensoredllm.ai/guides/cursor) (steps checked 2026-11-22): Paste the key as the OpenAI API key, override the OpenAI base URL, add an uncensored model id. Ask and Agent mode both work; the errors and their fixes.
- [Cline with an uncensored model](https://uncensoredllm.ai/guides/cline) (steps checked 2026-11-22): API Provider on OpenAI Compatible, the base URL, the key and an uncensored model id, then the Model Configuration values each build wants. One model for Plan and another for Act; the cost per agent turn and the errors.
- [Roo Code with an uncensored model](https://uncensoredllm.ai/guides/roo-code) (steps checked 2026-11-02): Set the OpenAI Compatible provider to the API: Base URL, API Key, Model, then the Model Configuration fields filled from the live catalog. One profile per mode, the cost per agent turn and the errors.
- [Kilo Code with an uncensored model](https://uncensoredllm.ai/guides/kilo-code) (steps checked 2026-11-22): Kilo Code takes a custom OpenAI-compatible provider in its Providers settings and in kilo.jsonc: the base URL, the key, uncensored model ids with their costs. Both the extension and the CLI, the cost per agent turn and the errors.
- [Continue with an uncensored model](https://uncensoredllm.ai/guides/continue) (steps checked 2026-11-02): One models entry in config.yaml: provider openai, the base URL, the key from a secret, tool_use for Agent mode. The same file drives the VS Code and JetBrains extensions and the cn CLI; the cost per agent turn and the errors.
- [Aider with an uncensored model](https://uncensoredllm.ai/guides/aider) (steps checked 2026-11-22): The base URL and key as environment variables or in .aider.conf.yml, the openai/ prefix on the model id, and the metadata file that ends the unknown-model warning. A weak model for commits, architect mode, the cost per edit turn and the errors.
- [Zed with an uncensored model](https://uncensoredllm.ai/guides/zed) (steps checked 2026-11-02): One openai_compatible block in Zed’s settings.json: the API URL, uncensored ids with their context windows and capabilities, the key in the keychain or an env var. The cost per agent turn and the errors.
- [LangChain and LangGraph with an uncensored model](https://uncensoredllm.ai/guides/langchain) (steps checked 2026-11-02): ChatOpenAI with base_url and api_key in Python, configuration.baseURL in JavaScript, an uncensored model id, and the same instance handed to create_agent for a tool-calling agent. Token counts on every message; the cost per agent turn and the errors.
- [LlamaIndex with an uncensored model](https://uncensoredllm.ai/guides/llamaindex) (steps checked 2026-11-22): OpenAILike with api_base, api_key, an uncensored model id and is_chat_model=True; OpenAILikeEmbedding for the index on the same key. A RAG query engine over your documents, answered by an abliterated build.
- [CrewAI with uncensored models](https://uncensoredllm.ai/guides/crewai) (steps checked 2026-11-22): One LLM object per agent: the openai/ prefix on a catalog id, base_url and api_key. Which builds take tool calls, what a crew run costs, and the errors a crew hits first.
- [Pydantic AI with an uncensored model](https://uncensoredllm.ai/guides/pydantic-ai) (steps checked 2026-11-22): OpenAIChatModel with OpenAICompatibleProvider pointed at this API: a typed agent on an uncensored build, structured output by tool call or response_format, and the model switched from config.
- [Dify with uncensored models](https://uncensoredllm.ai/guides/dify) (steps checked 2026-11-22): Add this API under the OpenAI-API-compatible model provider: endpoint URL, key, model name, context size and max tokens per model. Uncensored builds for Dify chatbots, agents and workflows, plus embeddings for knowledge bases.
- [DSPy with uncensored models](https://uncensoredllm.ai/guides/dspy) (steps checked 2026-11-22): dspy.LM with the openai/ prefix, api_base and model_type="chat": one prediction on a catalog id first, then an optimizer that multiplies the calls.
- [Haystack with uncensored models](https://uncensoredllm.ai/guides/haystack) (steps checked 2026-11-22): OpenAIChatGenerator with api_base_url, a Secret read from the environment and a catalog id: run the generator on its own, then drop it into the pipeline.
- [Semantic Kernel with uncensored models](https://uncensoredllm.ai/guides/semantic-kernel) (steps checked 2026-11-22): The .NET OpenAI connector with an endpoint Uri, behind the SKEXP0010 flag: modelId, apiKey and endpoint as three separate values, and a text call before any plugin.
- [Spring AI with uncensored models](https://uncensoredllm.ai/guides/spring-ai) (steps checked 2026-11-22): spring-ai-starter-model-openai pinned to 2.1.0: whether base-url carries /v1 (it does on 2.x, not on 1.x), the model property, and a key-free way to see the final URL.
- [AutoGen with uncensored models](https://uncensoredllm.ai/guides/autogen) (steps checked 2026-11-22): OpenAIChatCompletionClient with base_url and a model_info dict filled from the catalog, one AssistantAgent, one bounded task. AutoGen is archived; this setup targets the latest release, 0.7.5.
- [LangGraph with an uncensored model node](https://uncensoredllm.ai/guides/langgraph) (steps checked 2026-11-22): ChatOpenAI with base_url inside one graph node, messages kept by the add_messages reducer, START to END: prove the graph ends before you add tools.
- [LangChain4j with an uncensored chat model](https://uncensoredllm.ai/guides/langchain4j) (steps checked 2026-11-22): OpenAiChatModel with a baseUrl that ends in /v1, the key from the environment and a catalog id: one reply from a plain main method before AI Services or Spring.
- [smolagents with an uncensored tool-calling model](https://uncensoredllm.ai/guides/smolagents) (steps checked 2026-11-22): OpenAIModel with api_base, a ToolCallingAgent, one lookup tool over static data and a hard max_steps: the round trip to test before an agent gets real tools.
- [Test the UncensoredLLM API in Postman](https://uncensoredllm.ai/guides/postman) (steps checked 2026-11-22): An environment holding the base URL and a secret key variable, a catalog GET that costs nothing, then one chat POST that is a real, billed request.
- [Migrate a Venice API client to UncensoredLLM](https://uncensoredllm.ai/guides/migrate-from-venice-api) (steps checked 2026-11-22): Three settings change, a fourth needs an audit: the base URL, a new key, a destination id per model, and every venice_parameters field your code sends.
- [OpenCode with an uncensored model](https://uncensoredllm.ai/guides/opencode) (steps checked 2026-11-22): One provider block in opencode.json: the OpenAI-compatible package, the base URL, the key from an env var, uncensored ids with their names. Pick one with /models; the cost per agent turn and the errors.
- [Hermes Agent with an uncensored model](https://uncensoredllm.ai/guides/hermes-agent) (steps checked 2026-11-06): One providers entry in ~/.hermes/config.yaml: api, key_env, transport chat_completions and the uncensored ids. The key in .env, /model to switch, the cost of an agent turn and the errors.
- [OpenClaw with an uncensored model](https://uncensoredllm.ai/guides/openclaw) (steps checked 2026-11-06): One models.providers block in ~/.openclaw/openclaw.json: baseUrl, the key from an environment variable, api openai-completions and a models list of uncensored builds. The default model, the cost of a session and the errors.
- [Codex CLI with an uncensored model](https://uncensoredllm.ai/guides/codex-cli) (steps checked 2026-11-06): One [model_providers] table in ~/.codex/config.toml: base_url, env_key and wire_api responses, with an uncensored model id. A profile file per model, the /model menu pitfall, the cost of a task and the errors.
- [Uncensored coding AI in Claude Code, Cursor and Open WebUI](https://uncensoredllm.ai/guides/uncensored-coding-ai) (steps checked 2026-11-22): Uncensored coding AI set up as an agent: the code builds with the refusal removed, in Claude Code through a router, in Cursor and in Open WebUI, a coding day costed at today’s rates, and what the acceptable-use page still forbids.
- [Uncensored image and video API guide](https://uncensoredllm.ai/guides/media) (steps checked 2026-10-26): OpenAI’s image API for the image models and one generations endpoint for video: the request, the settings that set the price, polling from Python, what an image and a clip cost at the live price, and the refund rule.
