Capabilities Overview

ZeroTwo is built around six capability areas that cover the full range of AI-powered work — from multi-model chat and deep research to creative generation, app automation, persistent memory, and team workspaces. This page walks through each area so you know exactly what ZeroTwo can do.

​

AI Chat

The chat interface at /c is ZeroTwo’s primary workspace. Every conversation is powered by your choice of model from a catalog of 95+ options spanning 19 providers. What’s included:

• 95+ models from OpenAI, Anthropic, Google, Mistral, DeepSeek, Cohere, xAI, Perplexity, Qwen, Groq, OpenRouter, Kimi K2, Venice, TheSys, ZAI, Inception, ByteDance, and more
• Model switching — change models at any point during a conversation; the new model receives the full conversation history
• Canvas mode — open a split-screen editor (7 editor types) alongside your chat; the AI edits canvas content directly
• Agent mode — enable autonomous multi-step task execution using your connected apps as tools
• Voice — real-time WebRTC voice input and output with 10 voice options
• File uploads — attach PDFs, images, audio, video, spreadsheets, code files, Jupyter notebooks, and more (1 file/message on Free; up to 10 on Pro+)
• Streaming responses with full markdown rendering, code highlighting, and table support
• Shared links — publish any chat as a read-only link at /share/chat/:id

​

Prompt Bar Tools

The prompt bar at the bottom of the chat exposes tool pills for activating specific capabilities:

​

Research

ZeroTwo offers two complementary research tools: Web Search for quick, grounded answers and Deep Research for thorough, multi-source investigations. Web Search

• Searches the live web before generating a response
• 8 focus modes for targeted searching:
  • Auto — ZeroTwo picks the best approach for your query
  • Company — research a specific business or organization
  • Research — academic papers, studies, technical documentation
  • News — current events and recent articles
  • PDF — finds and reads PDF documents from the web
  • GitHub — searches code repositories, issues, and READMEs
  • Personal — queries your connected personal data sources
  • LinkedIn — searches professional profiles and company pages
• Sources appear in a right-panel Sources panel with titles, URLs, and snippets
• Works with any model that supports tool use

Deep Research

• Multi-step autonomous research powered by Firecrawl
• Workflow: ZeroTwo plans a research strategy → shows you the plan for approval → executes multi-step web browsing and synthesis → delivers a structured Canvas report with citations
• Reports include organized sections, key findings, and inline source references
• Available 1/day on Free; unlimited on Pro+

Real-time web data

• With Web Search or Perplexity Sonar models, responses are grounded in live web data
• Addresses the training cutoff problem — get current prices, news, and developments

​

Creative Studio

Dedicated studio pages for image, video, and audio generation keep creative work separate from your chat history with purpose-built controls. Image Generation (/studio/images)

• 10+ image generation models:
  • GPT-Image-1, GPT-Image-1.5, GPT-Image-mini (OpenAI) — photorealistic, strong instruction-following
  • Flux Pro v1.1, Flux Pro v2 (Black Forest Labs) — high-quality artistic outputs
  • Imagen 4.0 (Google) — photorealism and strong text rendering in images
  • Grok Imagine, Grok Imagine Pro (xAI) — creative and stylized generations
  • Qwen Image, Qwen Edit (Alibaba) — multilingual prompt support, image editing
  • LustIFY SDXL — SDXL-based generation
  • Klingai, Creatify, FAL-based models — specialized generation options
• Generation history with full-resolution outputs
• Available in chat via the Image tool pill or in the dedicated studio

Video Generation (/studio/video)

• Text-to-video and image-to-video generation
• Background job processing with results available in your files library
• Available on Pro+ plans

Audio Generation (/studio/audio)

• Generate speech, music, and sound effects from text prompts
• Automatic transcription — attach any audio file to a chat message and ZeroTwo transcribes it via Whisper; transcript is added to the conversation context
• Available on Pro+ plans

​

Connectors

ZeroTwo’s connectors bring live data from the apps you use into your AI conversations, and enable agent mode to take actions in those apps. 45+ OAuth integrations including: How connectors work:

• Connect via OAuth in Settings → Connectors (no API keys required)
• Use @mentions in chat to reference a specific connected app (@GitHub, @Notion, @GoogleDrive)
• In agent mode, ZeroTwo can take autonomous actions through connectors — create issues, update records, query databases, send emails
• Connector data is pulled live at the time of the request, not from training data

​

Memory & Personalization

ZeroTwo’s personalization layer ensures the AI always starts with context about you, without requiring you to re-introduce yourself in every conversation. Auto-learned memory

• ZeroTwo automatically learns facts about you from conversations — your role, preferences, technical stack, writing style, and more
• Memory entries are categorized as preferences, general facts, or profile information
• Free plan: 5 memory entries max · Pro+: unlimited
• Memories are injected automatically into the context of every new conversation

Memory management

• View and edit all memory entries at Settings → Personalization → Memory
• Delete individual entries or turn off auto-memory creation
• Export your memory as part of the full data export

Custom instructions

• Set in Settings → Personalization → Custom Instructions
• Two fields:
  • “What should ZeroTwo know about you?” — background, role, preferences
  • “How should ZeroTwo respond?” — tone, format, length, style
• Applied as system context to all new chats

Other personalization

• 11 languages — ZeroTwo supports 11 interface languages with RTL support for Arabic and Hebrew
• Light and dark themes — toggle in Settings
• Model preferences — set a default model for new conversations
• Voice preferences — choose from 10 voices for the voice interface

​

Projects & Teams

Projects give you organized, persistent workspaces for multi-session work. Custom agents and skills extend ZeroTwo’s automation capabilities. Business accounts add org-level management. Projects

• Group related chats, files, and instructions under a named project workspace
• Each project has its own instruction set (system prompt injected into every chat within the project)
• Project chats appear in a collapsible section in the sidebar, separate from standalone chat history
• Navigate directly to /projects/:id
• Free plan: 1 project · Pro+: unlimited

Custom agents (/agents)

• Build named agents with custom system prompts, tool configurations, and model assignments
• Agents can be given specific personas, capabilities, and connector access
• Reusable across conversations

Skills

• Reusable prompt templates and tool presets
• Attach to any agent or chat session
• Enables consistent workflows and reduces repeated prompt engineering

Tasks (/tasks)

• Background task runner for long-running agent jobs
• Track status, see outputs, and manage running tasks

Business workspace

• Shared org workspace for all members
• Admin panel at /admin/* — member management, usage visibility, role controls
• Per-seat billing via Stripe
• 14-day free trial with card validation

​

What Makes ZeroTwo Different

Model flexibility, no lock-in. With 95+ models from 19 providers, you are never dependent on a single vendor. When OpenAI ships a breakthrough or Anthropic releases a new Claude, it is available in ZeroTwo immediately. All-in-one — no context switching. Web search, canvas editing, voice, file uploads, agent mode, image generation — all accessible from the same prompt bar. You do not need a separate tool for each task. Integrations that use your real data. ZeroTwo’s connectors pull live data from the apps you actually use. Ask about a real Google Drive document, a real GitHub issue, or a real Supabase row — not training-time snapshots. Persistent memory. ZeroTwo builds a personal context layer from your conversations. Combined with project-level instructions, the AI understands your context before you type a word. Studio as a first-class workspace. Image, video, and audio generation live in purpose-built studio pages with generation history, output management, and model-specific controls — not just inline chat responses. Business-ready from the start. The Business plan adds org workspaces, admin controls, and per-seat billing without requiring an enterprise contract.