ComfyUI Node
RM-LLM 0.6.0
Twelve LLM API providers with model discovery, dynamic controls, optional media input, and server-side credentials.
RM-LLM 0.6.0
- image
- video
- agent_request
- response
- reasoning
- response_json
◄providerOpenRouter►
◄model_name►
◄endpointAuto►
◄api_key_envOPENROUTER_API_KEY►
◄credential_sourceEnvironment variable►
◄system_promptYou are a helpful assistant.►
◄user_promptPlease respond to the following request.►
◄parameters_json{}►
◄timeout_seconds300►
◄key_ticket►
◄console_outputfalse►
◄chat_template_kwargs.enable_thinking—►
◄chat_template_kwargs.preserve_thinking—►
◄chat_template_kwargs.clear_thinking—►
◄chat_template_kwargs.thinking_budget—►
◄chat_template_kwargs.date_string—►
◄cache_control—►
◄debug—►
◄frequency_penalty—►
◄image_config—►
◄logit_bias—►
◄logprobs—►
◄max_completion_tokens—►
◄max_tokens—►
◄metadata—►
◄min_p—►
◄modalities—►
◄parallel_tool_calls—►
◄plugins—►
◄prediction—►
◄presence_penalty—►
◄prompt_cache_key—►
◄prompt_cache_options—►
◄reasoning—►
◄reasoning_effort—►
◄repetition_penalty—►
◄response_format—►
◄seed—►
◄service_tier—►
◄session_id—►
◄stop—►
◄stop_server_tools_when—►
◄temperature—►
◄tool_choice—►
◄tools—►
◄top_a—►
◄top_k—►
◄top_logprobs—►
◄top_p—►
◄trace—►
◄user—►
◄include_reasoning—►
◄stop_token_ids—►
◄include_stop_str_in_output—►
◄min_tokens—►
◄chat_template_kwargs—►
◄n—►
◄output_config—►
◄thinking—►
◄thinking_config—►
◄safety_settings—►
◄creativity—►
◄reasoning.enabled—►
◄input_budget0►
◄thinking_budget0►
◄output_budget0►
CategoryRM/API
Inputs (69)
| Name | Type | Default | Description |
|---|---|---|---|
| provider | STRING | OpenRouter | — |
| model_name | STRING | Click model_name in the RM-LLM panel to fetch and search the complete provider catalog. | |
| endpoint | STRING | Auto | — |
| api_key_env | STRING | OPENROUTER_API_KEY | Environment variable NAME only. The server reads its value; it is never sent to the browser. |
| credential_source | STRING | Environment variable | — |
| system_prompt | STRING | You are a helpful assistant. | — |
| user_prompt | STRING | Please respond to the following request. | — |
| parameters_json | STRING | {} | — |
| timeout_seconds | INT | 30010–3600 | — |
| imageopt | IMAGE | — | |
| videoopt | VIDEO | — | |
| key_ticketopt | STRING | — | |
| console_outputopt | BOOLEAN | false | Stream generated text and returned reasoning live to the ComfyUI console. |
| chat_template_kwargs.enable_thinkingopt | BOOLEAN | Enable the model's thinking mode. | |
| chat_template_kwargs.preserve_thinkingopt | BOOLEAN | Preserve reasoning in supplied conversation history. | |
| chat_template_kwargs.clear_thinkingopt | BOOLEAN | Clear reasoning from supplied conversation history. | |
| chat_template_kwargs.thinking_budgetopt | INT | Requested reasoning token budget. | |
| chat_template_kwargs.date_stringopt | STRING | Date supplied to the chat template. | |
| cache_controlopt | STRING | Enable automatic prompt caching. When set at the top level, the system automatically applies cache breakpoints to the last cacheable block in the request. When set on an individual content block, it marks an explicit cache breakpoint; block-level markers also work on OpenAI models that support explicit prompt caching — OpenRouter converts them to the provider's native format. | |
| debugopt | STRING | Debug options for inspecting request transformations (streaming only) | |
| frequency_penaltyopt | FLOAT | Model setting. Structured values are supplied as JSON text. | |
| image_configopt | STRING | Provider-specific image configuration options. Keys and values vary by model/provider. See https://openrouter.ai/docs/guides/overview/multimodal/image-generation for more details. | |
| logit_biasopt | STRING | Token ID to bias mapping. | |
| logprobsopt | BOOLEAN | Return token log probabilities when supported. | |
| max_completion_tokensopt | INT | Model setting. Structured values are supplied as JSON text. | |
| max_tokensopt | INT | Model setting. Structured values are supplied as JSON text. | |
| metadataopt | STRING | Key-value pairs for additional object information (max 16 pairs, 64 char keys, 512 char values) | |
| min_popt | FLOAT | Optional. Sets a minimum probability threshold relative to the most likely token for a token to be considered. Must be between 0 and 1. Set to 0 to disable. | |
| modalitiesopt | STRING | Output modalities for the response. Supported values are "text", "image", and "audio". | |
| parallel_tool_callsopt | BOOLEAN | Whether to enable parallel function calling during tool use. When true, the model may generate multiple tool calls in a single response. | |
| pluginsopt | STRING | Plugins you want to enable for this request, including their settings. | |
| predictionopt | STRING | Static predicted output content. Supported models can use this to reduce latency when much of the response is known in advance. | |
| presence_penaltyopt | FLOAT | Model setting. Structured values are supplied as JSON text. | |
| prompt_cache_keyopt | STRING | Model setting. Structured values are supplied as JSON text. | |
| prompt_cache_optionsopt | STRING | Request-level prompt-cache controls. `mode: "explicit"` disables OpenAI-managed breakpoints so only blocks marked with `prompt_cache_breakpoint` are cached. Only supported by OpenAI GPT-5.6 and newer. | |
| reasoningopt | STRING | RM-LLM reasoning settings: enabled, effort or max_tokens, when supported. | |
| reasoning_effortopt | STRING | Model setting. Structured values are supplied as JSON text. | |
| repetition_penaltyopt | FLOAT | Optional. Penalizes new tokens based on their appearance in the prompt and generated text. Values > 1 encourage new tokens; < 1 encourages repetition. | |
| response_formatopt | STRING | Model setting. Structured values are supplied as JSON text. | |
| seedopt | INT | Model setting. Structured values are supplied as JSON text. | |
| service_tieropt | STRING | The service tier to use for processing this request. `fast` is accepted as an alias for `priority`. | |
| session_idopt | STRING | A unique identifier for grouping related requests (e.g., a conversation or agent workflow). When provided, OpenRouter uses it as the sticky routing key, routing all requests in the session to the same provider to maximize prompt cache hits. Also used for observability grouping. If provided in both the request body and the x-session-id header, the body value takes precedence. Maximum of 256 characters. | |
| stopopt | STRING | Model setting. Structured values are supplied as JSON text. | |
| stop_server_tools_whenopt | STRING | Stop conditions for the server-tool agent loop. Any condition firing halts the loop (OR logic). When set, this overrides `max_tool_calls`. When a condition fires while the model is still emitting tool calls, the pending tool calls are executed and one final turn is made with tool calls disabled so the response ends with a natural-language answer instead of an unfinished tool call. | |
| temperatureopt | FLOAT | Model setting. Structured values are supplied as JSON text. | |
| tool_choiceopt | STRING | Model setting. Structured values are supplied as JSON text. | |
| toolsopt | STRING | Model setting. Structured values are supplied as JSON text. | |
| top_aopt | FLOAT | Consider only tokens with "sufficiently high" probabilities based on the probability of the most likely token. Not all providers support this parameter. | |
| top_kopt | INT | Model setting. Structured values are supplied as JSON text. | |
| top_logprobsopt | INT | Number of alternative token log probabilities. | |
| top_popt | FLOAT | Model setting. Structured values are supplied as JSON text. | |
| traceopt | STRING | Metadata for observability and tracing. Known keys (trace_id, trace_name, span_name, generation_name, parent_span_id) have special handling. Additional keys are passed through as custom metadata to configured broadcast destinations. | |
| useropt | STRING | Optional end-user identifier. | |
| include_reasoningopt | BOOLEAN | Include reasoning in the response when supported. | |
| stop_token_idsopt | STRING | Optional. Similar to stop, but uses token IDs to halt generation. The output might include these tokens unless they are special tokens. | |
| include_stop_str_in_outputopt | BOOLEAN | Optional. If set to true, includes stop strings in the output text. Defaults to false. | |
| min_tokensopt | INT | Optional. The minimum number of tokens to generate before EOS or stop_token_ids can be generated. | |
| chat_template_kwargsopt | STRING | Optional. Model-specific chat template parameters, such as toggling reasoning. See the chat_template_kwargs guide. | |
| nopt | INT | Number of choices. LithosAI Kimi K3 requires 1. | |
| output_configopt | STRING | Anthropic output_config, including effort where supported. | |
| thinkingopt | STRING | Provider-native thinking settings. Do not combine with the Thinking/Reasoning toggle. | |
| thinking_configopt | STRING | Gemini thinkingConfig, using its native field names. | |
| safety_settingsopt | STRING | Gemini safetySettings, using its native field names. | |
| creativityopt | FLOAT | 0–100 | 0–100: sets Temperature and top_p. Individually connected sampling inputs take precedence. |
| reasoning.enabledopt | BOOLEAN | Enable reasoning when supported by the selected model. | |
| agent_requestopt | RM_LLM_AGENT_REQUEST | A connected tool conversation uses the Responses transport. System/user prompts and model settings remain on this RM-LLM node. | |
| input_budgetopt | INT | 00–2147483647 | Tokens; 0 keeps existing/default behaviour. Positive budgets take precedence over corresponding token settings. Thinking and output may share one API limit. |
| thinking_budgetopt | INT | 00–2147483647 | Tokens; 0 keeps existing/default behaviour. Positive budgets take precedence over corresponding token settings. Thinking and output may share one API limit. |
| output_budgetopt | INT | 00–2147483647 | Tokens; 0 keeps existing/default behaviour. Positive budgets take precedence over corresponding token settings. Thinking and output may share one API limit. |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| response | STRING | — |
| reasoning | STRING | — |
| response_json | STRING | — |