AI

AI integration · 12 node(s).

00Overview

Every field below is exactly what you see in the Flomation editor. Fields marked ● live picker let you choose from a list pulled live from your account — no IDs to look up.

01Connecting AI

Add your credentials as a Flomation environment secret, then pick them in each node. The connection fields are:

Field	Type		Details
API Key	`secret`	Required	sk-ant-...
Custom Endpoint	`string`		https://westeurope.api.cognitive.microsoft.com — overrides Resource Name
Base URL	`string`		http://ollama.internal:11434 — Azure OpenAI: https://my-resource.openai.azure.com
AWS Access Key ID	`secret`
AWS Secret Access Key	`secret`

Pick an Environment on your flow (Flow Settings → Environment) so the secret resolves. Secret fields never show the value — they reference ${secrets.your_secret}.

02Anthropic

Anthropic Prompt

ai/anthropic · Action

Send a prompt to the Anthropic Messages API and return the response

Field	Type		Details
Model	`secret`		choices: Claude Sonnet 4.6, Claude Haiku 4.5, Claude Opus 4.6
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Max Tokens	`integer`		2048
Temperature	`string`		0.7
Conversation History	`object`		${conversation_history}
Streaming (fires response per sentence for low-latency voice)	`boolean`
Tool Definitions (JSON)	`text`		[{"name":"web_search","description":"Search the web","input_schema":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}]

Returns: response, response_mode, should_respond, model, input_tokens, output_tokens, stop_reason, tool_calls_count, success, error

03Azure

Azure OpenAI Prompt

ai/azure_openai · Action

Send a prompt to an Azure OpenAI deployment and return the response

Field	Type		Details
Resource Name	`string`		my-resource — builds https://my-resource.openai.azure.com
Deployment	`string`	Required	The deployment name you chose in Azure, not a model id
API Version	`string`
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Max Tokens	`integer`		2048
Temperature	`string`		Leave blank for the model default — GPT-5 and o-series accept only 1
Top P	`string`		1.0
Frequency Penalty	`string`		0.0 (-2.0 to 2.0)
Presence Penalty	`string`		0.0 (-2.0 to 2.0)
Response Format	`string`		choices: Text, JSON, JSON Schema
JSON Schema	`text`		{"type":"object","properties":{"answer":{"type":"string"}},"required":["answer"],"additionalProperties":false}
Conversation History	`object`		${conversation_history}
Tool Definitions (JSON)	`text`		[{"type":"function","function":{"name":"web_search","description":"Search the web","parameters":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}}]

Returns: response, model, prompt_tokens, completion_tokens, total_tokens, should_respond, tool_calls_count, success, error

04Embed

Embed Text

ai/embed_text · Action

Turn text into an embedding — the numeric fingerprint of its meaning, for storing in or searching a vector database

Field	Type		Details
Embedding Provider	`string`	Required	choices: OpenAI, OpenAI-compatible (Azure, vLLM, LocalAI, TEI…), Azure OpenAI, Ollama (self-hosted), AWS Bedrock (Titan)
API Version	`string`		2024-10-21
AWS Region	`string`		us-east-1
Model	`combobox`	Required	text-embedding-3-small — choices: OpenAI text-embedding-3-small (1536 dimensions), OpenAI text-embedding-3-large (3072 dimensions), OpenAI text-embedding-ada-002 (1536 dimensions), Bedrock Titan Text v2 (1024 dimensions), Bedrock Titan Text v1 (1536 dimensions), Ollama nomic-embed-text (768 dimensions), Ollama mxbai-embed-large (1024 dimensions)
Text	`text`		The text to embed
Texts (JSON array)	`object`		Embed several at once: ["first", "second"] — overrides Text above
Dimensions	`integer`		Leave empty for the model's default. Must match your vector table (OpenAI 3-small = 1536, Bedrock Titan v2 = 1024)

Returns: embedding, embeddings, dimensions, count, model, result, tool_result, success, error

05Gemini

Gemini Prompt

ai/gemini · Action

Send a prompt to Google's Gemini API and return the response

Field	Type		Details
Model	`secret`		choices: Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.5 Flash Lite, Gemini 2.0 Flash, Gemini 2.0 Flash Lite
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Max Tokens	`integer`		8192
Temperature	`string`		0.7
Conversation History	`object`		${conversation_history}
Tool Definitions (JSON)	`text`		[{"name":"web_search","description":"Search the web","parameters":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}]

Returns: response, model, prompt_tokens, completion_tokens, total_tokens, should_respond, tool_calls_count, success, error

Gemini Image

ai/gemini_image · Action

Generate an image from a text prompt using Gemini 2.5 Flash Image

Field	Type		Details
Model	`secret`		choices: Gemini 2.5 Flash Image (Nano Banana), Gemini 2.0 Flash Preview Image Generation
Prompt	`text`	Required	A photorealistic image of a cat wearing a top hat, studio lighting.

Returns: tool_result, image_blob, image_base64, mime_type, model, success, error

Gemini Text-to-Speech

ai/gemini_tts · Action

Generate audio from a text prompt using Gemini's TTS preview

Field	Type		Details
Model	`secret`		choices: Gemini 2.5 Flash Preview TTS, Gemini 2.5 Pro Preview TTS
Text	`text`	Required	Say cheerfully: hello, world!
Voice	`secret`		choices: Kore (firm, default), Puck (upbeat), Charon (informative), Fenrir (excitable), Aoede (breezy), Leda (youthful), Orus (firm, deeper), Zephyr (bright)

Returns: tool_result, audio_blob, audio_base64, mime_type, voice, model, success, error

Gemini Video

ai/gemini_video · Action

Generate a short video clip from a text prompt using Google's Veo 2 or Veo 3

Field	Type		Details
Model	`secret`		choices: Veo 3.1 Fast (preview, with audio, cheapest), Veo 3.1 (preview, with audio, best quality), Veo 3.1 Lite (preview, with audio, lowest cost), Veo 2 (legacy, video only)
Prompt	`text`	Required	Photorealistic flythrough over snowy Welsh mountains at golden hour, drone footage.
Input Image (flo:blob token, optional — enables image-to-video)	`text`		flo:blob:abc...
Aspect ratio for the generated video. Set to '9:16' for portrait/vertical videos (TikTok, Instagram Reels, Instagram Stories, YouTube Shorts, mobile-first viewing). Set to '16:9' for landscape/horizontal (default, YouTube, cinematic). Veo does NOT support square (1:1). If the user mentions a vertical format, mobile-first, social shorts, or any portrait orientation, set this to '9:16'.	`text`		choices: 16:9 (landscape), 9:16 (portrait — TikTok / Reels / Stories)
Duration (seconds, 5–8 — model-dependent)	`integer`		8
Negative Prompt (things to avoid, optional)	`text`		blurry, low quality
Whether the generated video may include people. Set to 'allow_all' for any people (children + adults — Veo 2 only, downgraded to 'allow_adult' on Veo 3.1 in EU/UK/CH/MENA). Set to 'allow_adult' to allow adults only (the safest universal choice; works on every model and region). Set to 'dont_allow' to exclude people entirely (use for scenery, products, abstract motion). If the user mentions people, characters, faces, or human subjects in the prompt and doesn't object to people, set this to 'allow_adult'.	`text`		choices: Allow adults (default — works on every model and region), Allow all (Veo 2 only outside EU/UK; downgrades on Veo 3.1 in EU/UK), Don't allow people (scenery / products / abstract)
Seed (optional, for reproducibility)	`integer`		12345

Returns: tool_result, video_blob, video_base64, mime_type, model, operation_name, success, error

06Groq

Groq Prompt

ai/groq · Action

Send a prompt to the Groq Chat Completions API and return the response

Field	Type		Details
Model	`string`		llama-3.3-70b-versatile — choices: Llama 3.3 70B Versatile, Llama 3.1 8B Instant, Llama 4 Scout 17B, GPT-OSS 120B, GPT-OSS 20B, Qwen 3 32B, Groq Compound, Groq Compound Mini
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Max Tokens	`integer`		2048
Temperature	`string`		0.7
Conversation History	`object`		${conversation_history}
Tool Definitions (JSON)	`text`		[{"type":"function","function":{"name":"web_search","description":"Search the web","parameters":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}}]

Returns: response, model, prompt_tokens, completion_tokens, total_tokens, should_respond, tool_calls_count, success, error

07Ollama

Ollama Prompt

ai/ollama · Action

Send a prompt to a model running on your own Ollama server and return the response

Field	Type		Details
Model	`string`	Required · ● live picker	llama3.2 — choices: Llama 3.2, Llama 3.1 8B, Qwen 3, Gemma 3, Mistral, Phi-4, DeepSeek-R1, Qwen 2.5 Coder, Llama 3.2 Vision, LLaVA (Vision)
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Conversation History	`object`		${conversation_history}
Max Tokens	`integer`		1024 (-1 = unlimited)
Temperature	`string`		0.8
Response Format	`string`		choices: Text, JSON
Thinking	`string`		choices: Model Default, On, Off, On (Low Effort), On (Medium Effort), On (High Effort)
Advanced Options	`string`		choices: Hide, Show
Top P	`string`		0.9
Top K	`integer`		40
Min P	`string`		0.0
Seed	`integer`		Set for reproducible output
Stop Sequences	`string`		Comma-separated, e.g. ###,END
Keep Alive	`string`		5m
Context Length	`integer`		4096
Repetition Penalty	`string`		1.1 (1.0 = off)
Repeat Window	`integer`		64 (0 = off, -1 = context length)
Frequency Penalty	`string`		0.0
Presence Penalty	`string`		0.0
Prompt Batch Size	`integer`		512
GPU Layers	`integer`		-1 (auto)
Main GPU	`integer`		0
CPU Threads	`integer`		0 (auto)
Low VRAM Mode	`boolean`
Lock Model in RAM (mlock)	`boolean`
Memory-Map Model (mmap)	`boolean`
Penalize Newlines	`boolean`
Tool Definitions (JSON)	`text`		[{"type":"function","function":{"name":"web_search","description":"Search the web","parameters":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}}]

Returns: response, thinking, model, prompt_tokens, completion_tokens, total_tokens, should_respond, tool_calls_count, success, error

08Openai

OpenAI Prompt

ai/openai · Action

Send a prompt to the OpenAI Chat Completions API and return the response

Field	Type		Details
Model	`secret`		choices: GPT-4o, GPT-4o Mini, GPT-4.1, GPT-4.1 Mini, GPT-4.1 Nano, o3, o3 Mini, o4 Mini
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Max Tokens	`integer`		2048
Temperature	`string`		0.7
Conversation History	`object`		${conversation_history}
Tool Definitions (JSON)	`text`		[{"type":"function","function":{"name":"web_search","description":"Search the web","parameters":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}}]

Returns: response, model, prompt_tokens, completion_tokens, total_tokens, should_respond, tool_calls_count, success, error

09Openrouter

OpenRouter Prompt

ai/openrouter · Action

Send a prompt to any AI model through the OpenRouter unified API and return the response

Field	Type		Details
Model	`string`	● live picker	openai/gpt-5.4-mini — choices: OpenAI GPT-5.4, OpenAI GPT-5.4 Mini, Anthropic Claude Sonnet 5, Anthropic Claude Opus 4.8, Anthropic Claude Haiku 4.5, Google Gemini 3.1 Pro Preview, Google Gemini 2.5 Flash, Meta Llama 4 Maverick, DeepSeek V3.2, Qwen3 Max, Mistral Large 2512, GPT-OSS 20B (Free)
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Max Tokens	`integer`		2048
Temperature	`string`		0.7
Top P	`string`		1.0
Frequency Penalty	`string`		0.0 (-2.0 to 2.0)
Presence Penalty	`string`		0.0 (-2.0 to 2.0)
Response Format	`string`		choices: Text, JSON
Conversation History	`object`		${conversation_history}
Tool Definitions (JSON)	`text`		[{"type":"function","function":{"name":"web_search","description":"Search the web","parameters":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}}]

Returns: response, model, provider, prompt_tokens, completion_tokens, total_tokens, cost, should_respond, tool_calls_count, success, error

10Openwebui

Open WebUI Prompt

ai/openwebui · Action

Send a prompt to an Open WebUI (or any OpenAI-compatible) Chat Completions endpoint and return the response

Field	Type		Details
Model	`string`	Required	llama3.1:8b
System Prompt	`text`		You are a helpful assistant.
Prompt	`text`	Required	What would you like to ask?
Max Tokens	`integer`		2048
Temperature	`string`		0.7
Conversation History	`object`		${conversation_history}
Tool Definitions (JSON)	`text`		[{"type":"function","function":{"name":"web_search","description":"Search the web","parameters":{"type":"object","properties":{"query":{"type":"string"}},"required":["query"]}}}]

Returns: response, model, prompt_tokens, completion_tokens, total_tokens, should_respond, tool_calls_count, success, error

11Notes & Limitations

Behaviours and constraints worth knowing before you build with these nodes.

Conversation History is trimmed to fit the selected model's context window before each call, and the oldest messages are dropped first, so long histories can silently lose their earliest turns.
Top P and the frequency and presence penalties are sent only when you fill them in; left blank, they are omitted so the provider's own defaults apply rather than a neutral value.
Image attachments carried in the prompt are forwarded automatically to vision-capable models, whereas audio, video and document attachments are passed through only as plain-text references.
The Ollama and Open WebUI actions connect to the server URL you supply from wherever the Flomation runner executes, so a localhost or private-network address works only if the runner itself can reach that host.
Model is optional on the hosted-provider actions and falls back to a built-in default when left blank, while the Ollama, Open WebUI and Azure OpenAI actions require you to name the model or deployment explicitly.