One visual builder. All Foundation Models. Enterprise‑grade control.

ActionFlows connects to Amazon Bedrock, allowing you to orchestrate state-of-the-art models (text, vision, embeddings, and rerankers) with no code.

Why Amazon Bedrock on ActionFlows?

• Unified access to many models (Anthropic, Meta Llama, Mistral, Cohere, Stability, Amazon Nova/Titan) through a single node.
• No‑code workflow builder to chain prompts, tools, retrieval (RAG), and post‑processing.
• Enterprise security with AWS IAM/VPC patterns and fine‑grained secrets.
• Observability built‑in: request logs, latency, tokens, and errors.
• Scales with you: run flows on‑demand or via schedules, webhooks, and triggers.

Great for: chat assistants, knowledge search (RAG), summarization, content generation, ranking, image generation, and multimodal apps.

How It Works (4 steps)

1. Add “Amazon Bedrock” node in ActionFlows
2. Pick a model (e.g., Claude, Llama, Nova, Titan, Mistral, Cohere…)
3. Configure inputs (prompt, system, tools, retrieval, safety)
4. Run & monitor results (logs, retries, metrics)

Featured Capabilities

• Chat & Reasoning — high‑quality assistants with tool use and long‑context
• RAG — connect to your data sources; rerank and ground responses
• Vision & Images — analyze images or generate photorealistic assets
• Embeddings & Search — semantic search, clustering, retrieval
• Ranking — improve search relevance with rerankers for top‑k results

Supported Models (via Amazon Bedrock)

Use Search/Filter in the UI to explore by type (Text, Vision, Embeddings, Rerank, Multimodal). Model availability may vary by region and account.

AI21 Labs (Text)

• Jamba 1.5 Large — ai21.jamba-1-5-large-v1:0 — Advanced text generation & reasoning
• Jamba 1.5 Mini — ai21.jamba-1-5-mini-v1:0 — Fast, cost‑efficient text generation

Amazon Nova & Titan

• Nova Canvas — amazon.nova-canvas-v1:0 — Text→Image, high‑quality visuals
• Nova Lite — amazon.nova-lite-v1:0 — Lightweight multimodal, any‑to‑any
• Nova Micro — amazon.nova-micro-v1:0 — Ultra‑low‑latency text
• Nova Premier — amazon.nova-premier-v1:0 — Advanced multimodal (text/image/video)
• Nova Pro — amazon.nova-pro-v1:0 — Flagship multimodal for complex workflows
• Nova Reel — amazon.nova-reel-v1:0 — Creative text+image generation
• Nova Reel (v1:1) — amazon.nova-reel-v1:1 — Faster long‑form text variant
• Nova Sonic — amazon.nova-sonic-v1:0 — Ultra‑fast text for real‑time apps
• Rerank 1.0 — amazon.rerank-v1:0 — Search/result reranking
• Titan Image Generator G1 v2 — amazon.titan-image-generator-v2:0 — Photoreal text→image
• Titan Text G1 – Express — amazon.titan-text-express-v1 — Low‑latency text
• Titan Text G1 – Lite — amazon.titan-text-lite-v1 — Cost‑efficient text
• Titan Text G1 – Premier — amazon.titan-text-premier-v1:0 — Advanced text & reasoning
• Titan Embeddings G1 – Text — amazon.titan-embed-text-v1 — Text embeddings
• Titan Text Embeddings V2 — amazon.titan-embed-text-v2:0 — Improved multilingual embeddings
• Titan Multimodal Embeddings G1 — amazon.titan-embed-image-v1 — Unified text+image vectors

Anthropic Claude (Text / Multimodal)

• Claude 3 Haiku — anthropic.claude-3-haiku-20240307-v1:0 — Fast, cost‑efficient
• Claude 3 Sonnet — anthropic.claude-3-sonnet-20240229-v1:0 — Balanced enterprise model
• Claude 3 Opus — anthropic.claude-3-opus-20240229-v1:0 — Highest capability in 3‑series
• Claude 3.5 Haiku — anthropic.claude-3-5-haiku-20241022-v1:0 — Next‑gen speed + reasoning
• Claude 3.5 Sonnet — anthropic.claude-3-5-sonnet-20240620-v1:0 — Strong coding & reasoning
• Claude 3.5 Sonnet v2 — anthropic.claude-3-5-sonnet-20241022-v2:0 — Higher factuality, 128K ctx
• Claude 3.7 Sonnet — anthropic.claude-3-7-sonnet-20250219-v1:0 — Advanced reasoning (next evolution)
• Claude Opus 4 — anthropic.claude-opus-4-20250514-v1:0 — Frontier‑class hybrid reasoning
• Claude Sonnet 4 — anthropic.claude-sonnet-4-20250514-v1:0 — Balanced 4‑series with tool use

Cohere (Text / Embeddings / Rerank)

• Command Light — cohere.command-light-text-v14 — Fast, economical text
• Command — cohere.command-text-v14 — General‑purpose text generation
• Command R — cohere.command-r-v1:0 — RAG‑optimized, accurate Q&A
• Command R+ — cohere.command-r-plus-v1:0 — Advanced RAG with long context
• Embed English — cohere.embed-english-v3 — 1024‑d text embeddings
• Embed Multilingual — cohere.embed-multilingual-v3 — 100+ languages embeddings
• Rerank 3.5 — cohere.rerank-v3-5:0 — Improved ranking accuracy for RAG

DeepSeek (Text)

• DeepSeek‑R1 — deepseek.r1-v1:0 — Bilingual (EN/ZH), strong reasoning

Luma (Text)

• Ray v2 — luma.ray-v2:0 — Creative writing & character consistency

Meta Llama (Text)

• Llama 3 8B Instruct — meta.llama3-8b-instruct-v1:0 — Efficient, 8K ctx
• Llama 3 70B Instruct — meta.llama3-70b-instruct-v1:0 — Strong reasoning
• Llama 3.1 8B Instruct — meta.llama3-1-8b-instruct-v1:0 — Improved 8B reasoning
• Llama 3.1 70B Instruct — meta.llama3-1-70b-instruct-v1:0 — Better instruction following
• Llama 3.1 405B Instruct — meta.llama3-1-405b-instruct-v1:0 — Frontier‑scale
• Llama 3.2 1B / 3B / 11B / 90B Instruct — meta.llama3-2-…-instruct-v1:0 — Efficient→near‑frontier range
• Llama 3.3 70B Instruct — meta.llama3-3-70b-instruct-v1:0 — Faster inference, better tool use
• Llama 4 Maverick 17B Instruct — meta.llama4-maverick-17b-instruct-v1:0 — Real‑time tool use
• Llama 4 Scout 17B Instruct — meta.llama4-scout-17b-instruct-v1:0 — High‑speed inference
• Llama 4 Maverick 70B Instruct — meta.llama4-maverick-70b-instruct-v1:0 — SOTA reasoning

Mistral (Text / Multimodal)

• Mistral 7B Instruct — mistral.mistral-7b-instruct-v0:2 — Efficient 7B
• Mistral Small — mistral.mistral-small-2402-v1:0 — Cost‑effective production
• Mistral Large (Feb 2024) — mistral.mistral-large-2402-v1:0 — Multilingual, 32K ctx
• Mistral Large (Jul 2024) — mistral.mistral-large-2407-v1:0 — Enhanced reasoning
• Mixtral 8x7B Instruct — mistral.mixtral-8x7b-instruct-v0:1 — MoE, strong price/perf
• Pixtral Large — mistral.pixtral-large-2502-v1:0 — Vision‑language flagship

Stability AI (Images)

• Stable Diffusion 3.5 Large — stability.sd3-5-large-v1:0 — 1024×1024 photoreal
• Stable Image Core 1.0 — stability.stable-image-core-v1:1 — Efficient 768×768
• Stable Image Ultra 1.0 — stability.stable-image-ultra-v1:1 — Studio‑quality 1024×1024

Top Use Cases

• AI Chat Assistants — multilingual agents, tool use, long‑context Q&A
• RAG over private data — secure retrieval, rerank, grounded answers
• Document Summarization — contracts, tickets, medical/finance docs*
• Content Generation — product copy, blogs, ad variants, emails
• Search Relevance — rerankers + embeddings for superior results
• Vision & Creative — image understanding and high‑quality image generation

*Compliance requirements vary by industry; see Security & Governance below.

Security, Governance & Observability

• AWS IAM & Secrets — least‑privilege access, key rotation
• VPC‑friendly patterns — private connectivity where required
• Audit & Logs — request/response logging, error tracking, usage metrics
• PII controls — configurable redaction, prompt/response scrubbing
• Compliance‑friendly workflows — support for GDPR/SOC‑style controls (implementation‑dependent)

Pricing & Billing

ActionFlows usage is metered per run and storage. Model inference is billed by AWS Bedrock under your account according to each model’s pricing. Set quotas, rate limits, and alerts per project.