Caiioo

Tracking since Apr 1, 2026.

Changelog

• Jun 1, 2026
  description

  Caiioo — Browse, automate, write, code, and create with any AI model. Right from your sidebar.
  
  The agentic AI platform that lives in your browser and actually does things for you. Not just chat — it browses the web, drafts Gmail, builds slide decks, edits spreadsheets, generates images, writes code, and automates repetitive tasks. 500+ models. Zero vendor lock-in. Everything stays on your device.
  
  Free to use. No credit card required. Install and go in 60 seconds.
  
  
  WHAT YOU CAN DO
  
  🌐 Web Automation — Browse, search, click, fill forms, take screenshots, and wait for pages to load. Teach it about specific sites with Learned Pages so it gets smarter with every interaction.
  
  📝 Google Workspace — Read AND write Docs, Sheets, and Slides. Create meetings, draft emails, organize files — with just-in-time permissions so you control exactly what it can access.
  
  ✏️ Slate Editor — Built-in collaborative canvas for documents, code, spreadsheets, diagrams, and charts. Handles Word docs with tracked changes, PDFs, code in 20+ languages, and rich text. AI suggests edits with accept/reject workflow. Export to PDF, HTML, or Markdown.
  
  🎨 Image Generation — Create and edit images using Gemini, FLUX, Seedream, and more. Multiple models, aspect ratios, and resolutions. Results appear inline in your conversation.
  
  🎙️ Voice — Talk to Caiioo with speech-to-text (local Whisper or ElevenLabs). Have responses read aloud with natural voices. All local voice models run entirely on-device.
  
  📅 Calendar and Reminders — Manage events across Google Calendar and Apple Calendar. Get browser notifications for upcoming tasks.
  
  🤖 Custom Agents — Tell Caiioo "Make me an agent that..." and it builds one — a custom mode with its own system prompt, preferred tools, and personalized variables. Pre-built modes included. Save effective prompt patterns as reusable skills.
  
  🔍 Perplexity AI Search — Web-grounded search with citations, available with BYOK.
  
  🔌 API Integration — Call any REST API directly from conversations to fetch or send data.
  
  🧩 MCP Servers — Connect remote MCP servers and OpenAPI endpoints to extend Caiioo with unlimited tools.
  
  🛡️ Built-in Ad Blocking — Stop wasting AI tokens on ad content and illegible code. Powered by Ghostery.
  
  
  ANY AI MODEL — YOUR CHOICE
  
  Choose from 500+ models — Claude, GPT, Gemini, Llama, Mistral, and more. Switch models mid-conversation without losing context. New models available within hours of release. Or run fully offline with Ollama — your data never leaves your device.
  
  Caiioo Benchmarks built into the model selector gives you real-time performance data, cost comparisons, and capability rankings so you always pick the right model for the job.
  
  Get the full power of each AI. Caiioo sends your full context to the model — no hidden summarization, no forgetting. Better attention to detail, fewer hallucinations, and responses that actually use everything you gave them.
  
  
  WHY Caiioo
  
  ✅ 500+ AI Models — No vendor lock-in. Use any model. Switch mid-conversation.
  ✅ Full Context, No Throttling — Your entire conversation sent to the model. Nothing silently truncated.
  ✅ Privacy-First — All data stored locally. No Caiioo servers collect your data. No analytics. No telemetry. None.
  ✅ Google Workspace Write Access — Not just read. Actually create, edit, and manage your Docs, Sheets, Gmail, and Calendar.
  ✅ Built-in Editor — Slate gives you a full collaborative canvas. No other AI sidebar has this.
  ✅ Offline Mode — Run Ollama locally. Your data never leaves your device.
  ✅ MCP Extensibility — Connect any MCP server or OpenAPI endpoint for unlimited tool integrations.
  ✅ Works Everywhere — Chrome, Edge, Brave, Vivaldi, and more. No admin rights required.
  
  
  USE CASES
  
  🔹 Draft a reply to the email you're reading, using context from your Calendar and Drive
  🔹 Fill out a web form using data from a spreadsheet
  🔹 Research a topic across multiple sites and compile findings into a document
  🔹 Create calendar events from a list of dates on a webpage
  🔹 Generate an image for a presentation you're working on
  🔹 Describe how a web page works or summarize its content
  🔹 Build a custom agent for deal-hunting, workplace tasks, or any workflow
  🔹 Connect to n8n, Zapier, or any webhook via MCP or API integration
  🔹 Write, edit, and export documents without leaving your browser
  🔹 Run AI completely offline with local models
  
  
  HOW TO GET STARTED
  
  1. Install the extension
  2. Sign in and choose your AI setup — one-click add, paste your OpenRouter key, or set up Ollama
  3. Click the Caiioo icon to open the sidebar
  4. Start asking it to do things
  
  
  PRICING
  
  Free tier: Full AI chat with BYOK, web browsing, document reading, Perplexity search, and the Slate editor.
  
  Pro ($9/mo, 14-day free trial, no credit card): Full write access to Google Workspace, image generation, unlimited custom agents, private sync, remote MCP servers, and priority support.
  
  Built on patent-pending AI agent technology including dynamic per-turn tool selection, real-time streaming execution with mid-run interjection, and just-in-time context compression.
  
  Support: [email protected] | in-app ticketing

  Caiioo — Browse, automate, write, code, and create with any AI model. Right from your sidebar.
  
  The agentic AI platform that lives in your browser and actually does things for you. Not just chat — it browses the web, drafts Gmail, builds slide decks, edits spreadsheets, generates images, writes code, and automates repetitive tasks. 500+ models. Zero vendor lock-in. Everything stays on your device.
  
  Free to use. No credit card required. Install and go in 60 seconds.
  
  Security: CASA Tier 2 verified vendor, SOC 2 Type 1 audited.  All data except AI requests you make stay on device.  Supports Ollama for full on device privacy and Pseudonymizer available for cloud AI.
  
  
  WHAT YOU CAN DO
  
  🌐 Web Automation — Browse, search, click, fill forms, take screenshots, and wait for pages to load. Teach it about specific sites with Learned Pages so it gets smarter with every interaction.
  
  📝 Google Workspace — Read AND write Docs, Sheets, and Slides. Create meetings, draft emails, organize files — with just-in-time permissions so you control exactly what it can access.
  
  ✏️ Slate Editor — Built-in collaborative canvas for documents, code, spreadsheets, diagrams, and charts. Handles Word docs with tracked changes, PDFs, code in 20+ languages, and rich text. AI suggests edits with accept/reject workflow. Export to PDF, HTML, or Markdown.
  
  🎨 Image Generation — Create and edit images using Gemini, FLUX, Seedream, and more. Multiple models, aspect ratios, and resolutions. Results appear inline in your conversation.
  
  🎙️ Voice — Talk to Caiioo with speech-to-text (local Whisper or ElevenLabs). Have responses read aloud with natural voices. All local voice models run entirely on-device.
  
  📅 Calendar and Reminders — Manage events across Google Calendar and Apple Calendar. Get browser notifications for upcoming tasks.
  
  🤖 Custom Agents — Tell Caiioo "Make me an agent that..." and it builds one — a custom mode with its own system prompt, preferred tools, and personalized variables. Pre-built modes included. Save effective prompt patterns as reusable skills.
  
  🔍 Perplexity AI Search — Web-grounded search with citations, available with BYOK.
  
  🔌 API Integration — Call any REST API directly from conversations to fetch or send data.
  
  🧩 MCP Servers — Connect remote MCP servers and OpenAPI endpoints to extend Caiioo with unlimited tools.
  
  🛡️ Built-in Ad Blocking — Stop wasting AI tokens on ad content and illegible code. Powered by Ghostery.
  
  
  ANY AI MODEL — YOUR CHOICE
  
  Choose from 500+ models — Claude, GPT, Gemini, Llama, Mistral, and more. Switch models mid-conversation without losing context. New models available within hours of release. Or run fully offline with Ollama — your data never leaves your device.
  
  Caiioo Benchmarks built into the model selector gives you real-time performance data, cost comparisons, and capability rankings so you always pick the right model for the job.
  
  Get the full power of each AI. Caiioo sends your full context to the model — no hidden summarization, no forgetting. Better attention to detail, fewer hallucinations, and responses that actually use everything you gave them.
  
  
  WHY Caiioo
  
  ✅ 500+ AI Models — No vendor lock-in. Use any model. Switch mid-conversation.
  ✅ Full Context, No Throttling — Your entire conversation sent to the model. Nothing silently truncated.
  ✅ Privacy-First — All data stored locally. No Caiioo servers collect your data. No analytics. No telemetry. None.
  ✅ Google Workspace Write Access — Not just read. Actually create, edit, and manage your Docs, Sheets, Gmail, and Calendar.
  ✅ Built-in Editor — Slate gives you a full collaborative canvas. No other AI sidebar has this.
  ✅ Offline Mode — Run Ollama locally. Your data never leaves your device.
  ✅ MCP Extensibility — Connect any MCP server or OpenAPI endpoint for unlimited tool integrations.
  ✅ Works Everywhere — Chrome, Edge, Brave, Vivaldi, and more. No admin rights required.
  
  
  USE CASES
  
  🔹 Draft a reply to the email you're reading, using context from your Calendar and Drive
  🔹 Fill out a web form using data from a spreadsheet
  🔹 Research a topic across multiple sites and compile findings into a document
  🔹 Create calendar events from a list of dates on a webpage
  🔹 Generate an image for a presentation you're working on
  🔹 Describe how a web page works or summarize its content
  🔹 Build a custom agent for deal-hunting, workplace tasks, or any workflow
  🔹 Connect to n8n, Zapier, or any webhook via MCP or API integration
  🔹 Write, edit, and export documents without leaving your browser
  🔹 Run AI completely offline with local models
  
  
  HOW TO GET STARTED
  
  1. Install the extension
  2. Sign in and choose your AI setup — one-click add, paste your OpenRouter key, or set up Ollama
  3. Click the Caiioo icon to open the sidebar
  4. Start asking it to do things
  
  
  PRICING
  
  Free tier: Full AI chat with BYOK, web browsing, document reading, Perplexity search, and the Slate editor.
  
  Pro ($9/mo, 14-day free trial, no credit card): Full write access to Google Workspace, image generation, unlimited custom agents, private sync, remote MCP servers, and priority support.
  
  Built on patent-pending AI agent technology including dynamic per-turn tool selection, real-time streaming execution with mid-run interjection, and just-in-time context compression.
  
  Support: [email protected] | in-app ticketing

• Jun 1, 2026
  permissions

  sidePanel, storage, activeTab, scripting, tabs, tabGroups, unlimitedStorage, alarms, identity, declarativeNetRequest, declarativeNetRequestWithHostAccess, offscreen, notifications

  sidePanel, storage, activeTab, scripting, tabs, tabGroups, unlimitedStorage, alarms, identity, declarativeNetRequest, declarativeNetRequestWithHostAccess, offscreen, notifications, geolocation

• May 8, 2026
  description

  PebbleFlow — Browse, automate, write, code, and create with any AI model. Right from your sidebar.
  
  The agentic AI platform that lives in your browser and actually does things for you. Not just chat — it browses the web, drafts Gmail, builds slide decks, edits spreadsheets, generates images, writes code, and automates repetitive tasks. 500+ models. Zero vendor lock-in. Everything stays on your device.
  
  Free to use. No credit card required. Install and go in 60 seconds.
  
  
  WHAT YOU CAN DO
  
  🌐 Web Automation — Browse, search, click, fill forms, take screenshots, and wait for pages to load. Teach it about specific sites with Learned Pages so it gets smarter with every interaction.
  
  📝 Google Workspace — Read AND write Docs, Sheets, and Slides. Create meetings, draft emails, organize files — with just-in-time permissions so you control exactly what it can access.
  
  ✏️ Slate Editor — Built-in collaborative canvas for documents, code, spreadsheets, diagrams, and charts. Handles Word docs with tracked changes, PDFs, code in 20+ languages, and rich text. AI suggests edits with accept/reject workflow. Export to PDF, HTML, or Markdown.
  
  🎨 Image Generation — Create and edit images using Gemini, FLUX, Seedream, and more. Multiple models, aspect ratios, and resolutions. Results appear inline in your conversation.
  
  🎙️ Voice — Talk to PebbleFlow with speech-to-text (local Whisper or ElevenLabs). Have responses read aloud with natural voices. All local voice models run entirely on-device.
  
  📅 Calendar and Reminders — Manage events across Google Calendar and Apple Calendar. Get browser notifications for upcoming tasks.
  
  🤖 Custom Agents — Tell PebbleFlow "Make me an agent that..." and it builds one — a custom mode with its own system prompt, preferred tools, and personalized variables. Pre-built modes included. Save effective prompt patterns as reusable skills.
  
  🔍 Perplexity AI Search — Web-grounded search with citations, available with BYOK.
  
  🔌 API Integration — Call any REST API directly from conversations to fetch or send data.
  
  🧩 MCP Servers — Connect remote MCP servers and OpenAPI endpoints to extend PebbleFlow with unlimited tools.
  
  🛡️ Built-in Ad Blocking — Stop wasting AI tokens on ad content and illegible code. Powered by Ghostery.
  
  
  ANY AI MODEL — YOUR CHOICE
  
  Choose from 500+ models — Claude, GPT, Gemini, Llama, Mistral, and more. Switch models mid-conversation without losing context. New models available within hours of release. Or run fully offline with Ollama — your data never leaves your device.
  
  PebbleFlow Benchmarks built into the model selector gives you real-time performance data, cost comparisons, and capability rankings so you always pick the right model for the job.
  
  Get the full power of each AI. PebbleFlow sends your full context to the model — no hidden summarization, no forgetting. Better attention to detail, fewer hallucinations, and responses that actually use everything you gave them.
  
  
  WHY PEBBLEFLOW
  
  ✅ 500+ AI Models — No vendor lock-in. Use any model. Switch mid-conversation.
  ✅ Full Context, No Throttling — Your entire conversation sent to the model. Nothing silently truncated.
  ✅ Privacy-First — All data stored locally. No PebbleFlow servers collect your data. No analytics. No telemetry. None.
  ✅ Google Workspace Write Access — Not just read. Actually create, edit, and manage your Docs, Sheets, Gmail, and Calendar.
  ✅ Built-in Editor — Slate gives you a full collaborative canvas. No other AI sidebar has this.
  ✅ Offline Mode — Run Ollama locally. Your data never leaves your device.
  ✅ MCP Extensibility — Connect any MCP server or OpenAPI endpoint for unlimited tool integrations.
  ✅ Works Everywhere — Chrome, Edge, Brave, Vivaldi, and more. No admin rights required.
  
  
  USE CASES
  
  🔹 Draft a reply to the email you're reading, using context from your Calendar and Drive
  🔹 Fill out a web form using data from a spreadsheet
  🔹 Research a topic across multiple sites and compile findings into a document
  🔹 Create calendar events from a list of dates on a webpage
  🔹 Generate an image for a presentation you're working on
  🔹 Describe how a web page works or summarize its content
  🔹 Build a custom agent for deal-hunting, workplace tasks, or any workflow
  🔹 Connect to n8n, Zapier, or any webhook via MCP or API integration
  🔹 Write, edit, and export documents without leaving your browser
  🔹 Run AI completely offline with local models
  
  
  HOW TO GET STARTED
  
  1. Install the extension
  2. Sign in and choose your AI setup — one-click add, paste your OpenRouter key, or set up Ollama
  3. Click the PebbleFlow icon to open the sidebar
  4. Start asking it to do things
  
  
  PRICING
  
  Free tier: Full AI chat with BYOK, web browsing, document reading, Perplexity search, and the Slate editor.
  
  Pro ($9/mo, 14-day free trial, no credit card): Full write access to Google Workspace, image generation, unlimited custom agents, private sync, remote MCP servers, and priority support.
  
  Built on patent-pending AI agent technology including dynamic per-turn tool selection, real-time streaming execution with mid-run interjection, and just-in-time context compression.
  
  Support: [email protected] | in-app ticketing

  Caiioo — Browse, automate, write, code, and create with any AI model. Right from your sidebar.
  
  The agentic AI platform that lives in your browser and actually does things for you. Not just chat — it browses the web, drafts Gmail, builds slide decks, edits spreadsheets, generates images, writes code, and automates repetitive tasks. 500+ models. Zero vendor lock-in. Everything stays on your device.
  
  Free to use. No credit card required. Install and go in 60 seconds.
  
  
  WHAT YOU CAN DO
  
  🌐 Web Automation — Browse, search, click, fill forms, take screenshots, and wait for pages to load. Teach it about specific sites with Learned Pages so it gets smarter with every interaction.
  
  📝 Google Workspace — Read AND write Docs, Sheets, and Slides. Create meetings, draft emails, organize files — with just-in-time permissions so you control exactly what it can access.
  
  ✏️ Slate Editor — Built-in collaborative canvas for documents, code, spreadsheets, diagrams, and charts. Handles Word docs with tracked changes, PDFs, code in 20+ languages, and rich text. AI suggests edits with accept/reject workflow. Export to PDF, HTML, or Markdown.
  
  🎨 Image Generation — Create and edit images using Gemini, FLUX, Seedream, and more. Multiple models, aspect ratios, and resolutions. Results appear inline in your conversation.
  
  🎙️ Voice — Talk to Caiioo with speech-to-text (local Whisper or ElevenLabs). Have responses read aloud with natural voices. All local voice models run entirely on-device.
  
  📅 Calendar and Reminders — Manage events across Google Calendar and Apple Calendar. Get browser notifications for upcoming tasks.
  
  🤖 Custom Agents — Tell Caiioo "Make me an agent that..." and it builds one — a custom mode with its own system prompt, preferred tools, and personalized variables. Pre-built modes included. Save effective prompt patterns as reusable skills.
  
  🔍 Perplexity AI Search — Web-grounded search with citations, available with BYOK.
  
  🔌 API Integration — Call any REST API directly from conversations to fetch or send data.
  
  🧩 MCP Servers — Connect remote MCP servers and OpenAPI endpoints to extend Caiioo with unlimited tools.
  
  🛡️ Built-in Ad Blocking — Stop wasting AI tokens on ad content and illegible code. Powered by Ghostery.
  
  
  ANY AI MODEL — YOUR CHOICE
  
  Choose from 500+ models — Claude, GPT, Gemini, Llama, Mistral, and more. Switch models mid-conversation without losing context. New models available within hours of release. Or run fully offline with Ollama — your data never leaves your device.
  
  Caiioo Benchmarks built into the model selector gives you real-time performance data, cost comparisons, and capability rankings so you always pick the right model for the job.
  
  Get the full power of each AI. Caiioo sends your full context to the model — no hidden summarization, no forgetting. Better attention to detail, fewer hallucinations, and responses that actually use everything you gave them.
  
  
  WHY Caiioo
  
  ✅ 500+ AI Models — No vendor lock-in. Use any model. Switch mid-conversation.
  ✅ Full Context, No Throttling — Your entire conversation sent to the model. Nothing silently truncated.
  ✅ Privacy-First — All data stored locally. No Caiioo servers collect your data. No analytics. No telemetry. None.
  ✅ Google Workspace Write Access — Not just read. Actually create, edit, and manage your Docs, Sheets, Gmail, and Calendar.
  ✅ Built-in Editor — Slate gives you a full collaborative canvas. No other AI sidebar has this.
  ✅ Offline Mode — Run Ollama locally. Your data never leaves your device.
  ✅ MCP Extensibility — Connect any MCP server or OpenAPI endpoint for unlimited tool integrations.
  ✅ Works Everywhere — Chrome, Edge, Brave, Vivaldi, and more. No admin rights required.
  
  
  USE CASES
  
  🔹 Draft a reply to the email you're reading, using context from your Calendar and Drive
  🔹 Fill out a web form using data from a spreadsheet
  🔹 Research a topic across multiple sites and compile findings into a document
  🔹 Create calendar events from a list of dates on a webpage
  🔹 Generate an image for a presentation you're working on
  🔹 Describe how a web page works or summarize its content
  🔹 Build a custom agent for deal-hunting, workplace tasks, or any workflow
  🔹 Connect to n8n, Zapier, or any webhook via MCP or API integration
  🔹 Write, edit, and export documents without leaving your browser
  🔹 Run AI completely offline with local models
  
  
  HOW TO GET STARTED
  
  1. Install the extension
  2. Sign in and choose your AI setup — one-click add, paste your OpenRouter key, or set up Ollama
  3. Click the Caiioo icon to open the sidebar
  4. Start asking it to do things
  
  
  PRICING
  
  Free tier: Full AI chat with BYOK, web browsing, document reading, Perplexity search, and the Slate editor.
  
  Pro ($9/mo, 14-day free trial, no credit card): Full write access to Google Workspace, image generation, unlimited custom agents, private sync, remote MCP servers, and priority support.
  
  Built on patent-pending AI agent technology including dynamic per-turn tool selection, real-time streaming execution with mid-run interjection, and just-in-time context compression.
  
  Support: [email protected] | in-app ticketing

• May 8, 2026
  name

  PebbleFlow

  Caiioo

• Apr 8, 2026
  description

  AI-powered context management assistant with web browsing, file operations, and API integrations
  
  PebbleFlow — Browse, automate, write, code, and create with any AI model. Right from your sidebar.
  
  The agentic AI platform that lives in your browser and actually does things for you. Not just chat — it browses the web, drafts Gmail, builds slide decks, edits spreadsheets, generates images, writes code, and automates repetitive tasks. 500+ models. Zero vendor lock-in. Everything stays on your device.
  
  Free to use. No credit card required. Install and go in 60 seconds.
  
  
  WHAT YOU CAN DO
  
  🌐 Web Automation — Browse, search, click, fill forms, take screenshots, and wait for pages to load. Teach it about specific sites with Learned Pages so it gets smarter with every interaction.
  
  📝 Google Workspace — Read AND write Docs, Sheets, Slides, Gmail, Calendar, and Drive. Create meetings, draft emails, organize files — with just-in-time permissions so you control exactly what it can access.
  
  ✏️ Slate Editor — Built-in collaborative canvas for documents, code, spreadsheets, diagrams, and charts. Handles Word docs with tracked changes, PDFs, code in 20+ languages, and rich text. AI suggests edits with accept/reject workflow. Export to PDF, HTML, or Markdown.
  
  🎨 Image Generation — Create and edit images using Gemini, FLUX, Seedream, and more. Multiple models, aspect ratios, and resolutions. Results appear inline in your conversation.
  
  🎙️ Voice — Talk to PebbleFlow with speech-to-text (local Whisper or ElevenLabs). Have responses read aloud with natural voices. All local voice models run entirely on-device.
  
  📅 Calendar and Reminders — Manage events across Google Calendar and Apple Calendar. Get browser notifications for upcoming tasks.
  
  🤖 Custom Agents — Tell PebbleFlow "Make me an agent that..." and it builds one — a custom mode with its own system prompt, preferred tools, and personalized variables. Pre-built modes included. Save effective prompt patterns as reusable skills.
  
  🔍 Perplexity AI Search — Web-grounded search with citations, available with BYOK.
  
  🔌 API Integration — Call any REST API directly from conversations to fetch or send data.
  
  🧩 MCP Servers — Connect remote MCP servers and OpenAPI endpoints to extend PebbleFlow with unlimited tools.
  
  🛡️ Built-in Ad Blocking — Stop wasting AI tokens on ad content and illegible code. Powered by Ghostery.
  
  
  ANY AI MODEL — YOUR CHOICE
  
  Choose from 500+ models — Claude, GPT, Gemini, Llama, Mistral, and more. Switch models mid-conversation without losing context. New models available within hours of release. Or run fully offline with Ollama — your data never leaves your device.
  
  PebbleFlow Benchmarks built into the model selector gives you real-time performance data, cost comparisons, and capability rankings so you always pick the right model for the job.
  
  Get the full power of each AI. PebbleFlow sends your full context to the model — no hidden summarization, no forgetting. Better attention to detail, fewer hallucinations, and responses that actually use everything you gave them.
  
  
  WHY PEBBLEFLOW
  
  ✅ 500+ AI Models — No vendor lock-in. Use any model. Switch mid-conversation.
  ✅ Full Context, No Throttling — Your entire conversation sent to the model. Nothing silently truncated.
  ✅ Privacy-First — All data stored locally. No PebbleFlow servers collect your data. No analytics. No telemetry. None.
  ✅ Google Workspace Write Access — Not just read. Actually create, edit, and manage your Docs, Sheets, Gmail, and Calendar.
  ✅ Built-in Editor — Slate gives you a full collaborative canvas. No other AI sidebar has this.
  ✅ Offline Mode — Run Ollama locally. Your data never leaves your device.
  ✅ MCP Extensibility — Connect any MCP server or OpenAPI endpoint for unlimited tool integrations.
  ✅ Works Everywhere — Chrome, Edge, Brave, Vivaldi, and more. No admin rights required.
  
  
  USE CASES
  
  🔹 Draft a reply to the email you're reading, using context from your Calendar and Drive
  🔹 Fill out a web form using data from a spreadsheet
  🔹 Research a topic across multiple sites and compile findings into a document
  🔹 Create calendar events from a list of dates on a webpage
  🔹 Generate an image for a presentation you're working on
  🔹 Describe how a web page works or summarize its content
  🔹 Build a custom agent for deal-hunting, workplace tasks, or any workflow
  🔹 Connect to n8n, Zapier, or any webhook via MCP or API integration
  🔹 Write, edit, and export documents without leaving your browser
  🔹 Run AI completely offline with local models
  
  
  HOW TO GET STARTED
  
  1. Install the extension
  2. Sign in and choose your AI setup — one-click add, paste your OpenRouter key, or set up Ollama
  3. Click the PebbleFlow icon to open the sidebar
  4. Start asking it to do things
  
  
  PRICING
  
  Free tier: Full AI chat with BYOK, web browsing, document reading, Perplexity search, and the Slate editor.
  
  Pro ($9/mo, 14-day free trial, no credit card): Full write access to Google Workspace, image generation, unlimited custom agents, private sync, remote MCP servers, and priority support.
  
  Built on patent-pending AI agent technology including dynamic per-turn tool selection, real-time streaming execution with mid-run interjection, and just-in-time context compression.
  
  Support: [email protected] | in-app ticketing

  PebbleFlow — Browse, automate, write, code, and create with any AI model. Right from your sidebar.
  
  The agentic AI platform that lives in your browser and actually does things for you. Not just chat — it browses the web, drafts Gmail, builds slide decks, edits spreadsheets, generates images, writes code, and automates repetitive tasks. 500+ models. Zero vendor lock-in. Everything stays on your device.
  
  Free to use. No credit card required. Install and go in 60 seconds.
  
  
  WHAT YOU CAN DO
  
  🌐 Web Automation — Browse, search, click, fill forms, take screenshots, and wait for pages to load. Teach it about specific sites with Learned Pages so it gets smarter with every interaction.
  
  📝 Google Workspace — Read AND write Docs, Sheets, and Slides. Create meetings, draft emails, organize files — with just-in-time permissions so you control exactly what it can access.
  
  ✏️ Slate Editor — Built-in collaborative canvas for documents, code, spreadsheets, diagrams, and charts. Handles Word docs with tracked changes, PDFs, code in 20+ languages, and rich text. AI suggests edits with accept/reject workflow. Export to PDF, HTML, or Markdown.
  
  🎨 Image Generation — Create and edit images using Gemini, FLUX, Seedream, and more. Multiple models, aspect ratios, and resolutions. Results appear inline in your conversation.
  
  🎙️ Voice — Talk to PebbleFlow with speech-to-text (local Whisper or ElevenLabs). Have responses read aloud with natural voices. All local voice models run entirely on-device.
  
  📅 Calendar and Reminders — Manage events across Google Calendar and Apple Calendar. Get browser notifications for upcoming tasks.
  
  🤖 Custom Agents — Tell PebbleFlow "Make me an agent that..." and it builds one — a custom mode with its own system prompt, preferred tools, and personalized variables. Pre-built modes included. Save effective prompt patterns as reusable skills.
  
  🔍 Perplexity AI Search — Web-grounded search with citations, available with BYOK.
  
  🔌 API Integration — Call any REST API directly from conversations to fetch or send data.
  
  🧩 MCP Servers — Connect remote MCP servers and OpenAPI endpoints to extend PebbleFlow with unlimited tools.
  
  🛡️ Built-in Ad Blocking — Stop wasting AI tokens on ad content and illegible code. Powered by Ghostery.
  
  
  ANY AI MODEL — YOUR CHOICE
  
  Choose from 500+ models — Claude, GPT, Gemini, Llama, Mistral, and more. Switch models mid-conversation without losing context. New models available within hours of release. Or run fully offline with Ollama — your data never leaves your device.
  
  PebbleFlow Benchmarks built into the model selector gives you real-time performance data, cost comparisons, and capability rankings so you always pick the right model for the job.
  
  Get the full power of each AI. PebbleFlow sends your full context to the model — no hidden summarization, no forgetting. Better attention to detail, fewer hallucinations, and responses that actually use everything you gave them.
  
  
  WHY PEBBLEFLOW
  
  ✅ 500+ AI Models — No vendor lock-in. Use any model. Switch mid-conversation.
  ✅ Full Context, No Throttling — Your entire conversation sent to the model. Nothing silently truncated.
  ✅ Privacy-First — All data stored locally. No PebbleFlow servers collect your data. No analytics. No telemetry. None.
  ✅ Google Workspace Write Access — Not just read. Actually create, edit, and manage your Docs, Sheets, Gmail, and Calendar.
  ✅ Built-in Editor — Slate gives you a full collaborative canvas. No other AI sidebar has this.
  ✅ Offline Mode — Run Ollama locally. Your data never leaves your device.
  ✅ MCP Extensibility — Connect any MCP server or OpenAPI endpoint for unlimited tool integrations.
  ✅ Works Everywhere — Chrome, Edge, Brave, Vivaldi, and more. No admin rights required.
  
  
  USE CASES
  
  🔹 Draft a reply to the email you're reading, using context from your Calendar and Drive
  🔹 Fill out a web form using data from a spreadsheet
  🔹 Research a topic across multiple sites and compile findings into a document
  🔹 Create calendar events from a list of dates on a webpage
  🔹 Generate an image for a presentation you're working on
  🔹 Describe how a web page works or summarize its content
  🔹 Build a custom agent for deal-hunting, workplace tasks, or any workflow
  🔹 Connect to n8n, Zapier, or any webhook via MCP or API integration
  🔹 Write, edit, and export documents without leaving your browser
  🔹 Run AI completely offline with local models
  
  
  HOW TO GET STARTED
  
  1. Install the extension
  2. Sign in and choose your AI setup — one-click add, paste your OpenRouter key, or set up Ollama
  3. Click the PebbleFlow icon to open the sidebar
  4. Start asking it to do things
  
  
  PRICING
  
  Free tier: Full AI chat with BYOK, web browsing, document reading, Perplexity search, and the Slate editor.
  
  Pro ($9/mo, 14-day free trial, no credit card): Full write access to Google Workspace, image generation, unlimited custom agents, private sync, remote MCP servers, and priority support.
  
  Built on patent-pending AI agent technology including dynamic per-turn tool selection, real-time streaming execution with mid-run interjection, and just-in-time context compression.
  
  Support: [email protected] | in-app ticketing

Permissions & access

Permissions

sidePanelstorageactiveTabscriptingtabstabGroupsunlimitedStoragealarmsidentitydeclarativeNetRequestdeclarativeNetRequestWithHostAccessoffscreennotificationsgeolocation

Host access

<all_urls>

Screenshots

Similar extensions

Alternatives to Caiioo, ranked by description similarity.

Prompt Anything Pro - Select Text & Prompt Any AI

Select any text, prompt AI instantly. BYOK (OpenAI, Anthropic, Google). Save prompts, build your library. Privacy-first.

19

★ 5.0

Vex 🍒

Your AI-Powered Life

3

Universal AI Navigator

Navigation sidebar for Gemini, ChatGPT, Claude, and Perplexity.

18

★ 5.0

Bibrowse

AI agent that automates any cross-site browser tasks for you. Why click when you can just ask?

24

★ 5.0

Tocify.ai – ChatGPT & Claude to Structured Word Docs

Generate a Table of Contents from ChatGPT, Claude, Gemini & Perplexity chats. One-click export and share-link detection.

10

Your AI Personal Assistant for Google Workspace.

AIPA drafts emails, schedules meetings, adds tasks and summarizes any page from a tiny assistant in Chrome who's always on your tab.

162

★ 5.0

Promptly - Enhance your prompts!

Upgrade everyday prompts into polished, expert-level instructions using AI

1

AI Chat Organizer

Organize your AI conversations with folders, tags, and powerful search. Works with ChatGPT and Claude. Your data stays local.

18

★ 5.0