The AI
you need.
Choose your models. Bring your files, tools and ideas into one conversation. Connect OpenRouter and pay only for the model usage you choose.
Your AI.
Your choice.
Connect cloud models through OpenRouter, or bring Codex and Pi on your Mac into NanthAI. Choose the connection that suits your work.
Connect the AI you use
Create a free NanthAI account and connect OpenRouter for cloud models. With Pro, you can also pair your Mac to use Codex or Pi.
Choose who does the work
Pick a cloud model for a quick answer, or a Local Persona to work with files on your Mac. Bring another model into the conversation when you want a second opinion.
Choose how you pay for AI
Use OpenRouter credit, your Codex subscription, or the provider you configure in Pi. Pi can also run models installed on your Mac through Ollama.
analysis
draft
cross-check
From a question to something you can use.
Compare three models against a source-backed brief, attach the output to a scheduled job, then keep the useful branches as files or an ideascape.
Files and knowledge base
sources + citations
Search and research
web research
Scheduled jobs
weekly job
One app,
full workflow.
Keep your ideas, files and AI in one place. Add Personas, memory and scheduled tasks, or connect your Mac to keep work moving from your phone.
Goals in Collaboration
Add a goal and describe a successful result. Let the team choose an owner or assign one yourself; reviewers and dependencies are optional. Cloud and supported local runtimes use the same workflow.
Learn moreContinue a conversation in Telegram
Continue your NanthAI chat with the same AIs and conversation mode. Send text, photos, files and voice; voice transcription and spoken replies need OpenRouter and follow your audio settings. Link your own account in Settings. Local Personas need the paired Mac online; approvals stay in signed-in NanthAI.
Learn moreLocal Runtime
Work with files on your Mac using Codex or Pi, and continue from your phone or browser. Local runtime is included with Pro.
Learn moreMulti-model chat
Add up to three models to one conversation. Compare answers side by side, fork a message to explore a different direction, or regenerate with a different model. Queue follow-ups while the AI is still responding.
Custom personas
Create a Persona once, then use it as your chat participant, a private Advisor, a scheduled-task specialist, or part of a multi-model discussion.
Persistent memory
Edge can remember facts across conversations — your preferences, projects, writing style, relationships, and goals. Memory is categorised, searchable, fully visible, and explorable as an interactive relationship graph. Set it to save automatically, ask first, or stay off entirely.
Search and research
Internet search is available on the free tier. Pro unlocks deeper modes: web search at three complexity levels, and a full research pipeline that plans queries, iterates on sources, and synthesises a cited paper. Attach results to any conversation.
Scheduled jobs
Set recurring AI tasks — daily briefings, weekly reports, monitoring jobs — and let them run in the background. Each job can have up to five steps with their own model, persona, search mode, and integrations. Results appear as regular conversations.
Autonomous sessions
Let multiple personas debate, brainstorm, or review each other's work without your input. Configure the number of cycles, set a moderator, enable stop-on-consensus, and intervene at any point. Cost estimates shown before you start.
Files and knowledge base
Attach images, PDFs, and Office documents to any chat. Pro adds a persistent knowledge base — upload reference material once and pull it into conversations or scheduled jobs whenever you need it. Edge can also generate documents, spreadsheets, and presentations.
Document workflows
Create and refine Word documents, Excel workbooks, and PowerPoint presentations from chat. Open or download finished files, reuse them later, and keep source material and revisions with the conversation.
Learn moreAI Skills & focused helpers
Choose from 60+ built-in skills or create your own specialist instructions. For larger tasks, eligible Pro chats can delegate independent work to up to three focused helpers and combine the results.
Learn moreAnalysis, code & charts
Ask Edge to inspect uploaded data, run calculations in an isolated workspace, create charts, and return reusable files. Standard Python analysis is included with Pro, with a bounded cloud sandbox for heavier work.
Learn moreVoice messages
Record up to two minutes, review the system transcript, and send it to any model as text. Listen to saved recordings or generate a read-aloud version of assistant replies using your OpenRouter balance.
Image generation
Pick an image-generation model, describe what you want, and receive an inline image. Download, share, or attach it to another conversation. Generated images are saved to the Knowledge Base.
Video generation
Choose a video model like Veo, Seedance, or Wan. Set resolution, duration, and aspect ratio, optionally attach first and last frame images, and get an inline video player with download and fullscreen.
Audio generation
Select a music model like Google Lyria, type a prompt, and get an MP3 back. Play, seek, adjust speed, and download — all without leaving the conversation. Generated audio appears in the Knowledge Base.
Voice references per request
Choose a compatible speech model, attach or record 1–3 permitted voice references, and confirm permission for that request. Required transcripts and clip counts depend on the selected model. References do not become a global voice.
Learn moreIdeascapes
Switch any conversation into a spatial canvas. Arrange AI responses as movable nodes, draw non-linear paths between ideas, and zoom across the full map. Positions persist — come back to the same layout later.
Connected
to your stack.
Connect your existing accounts and give Edge access to the services you already depend on. Every integration works inside regular chats, persona conversations, and scheduled jobs.
Google Drive & Calendar
Upload and browse Drive files, view and create calendar events, and use Gmail separately through Manual Gmail. All accessible from chat, personas, and scheduled jobs.
Microsoft 365
Same capabilities as Google — email, file storage, and calendar management — for Microsoft accounts. Manage folders, move messages, create events.
Notion
Search your workspace, read and create pages, query and update databases. Use Notion as a living knowledge source inside any conversation.
Apple Calendar
Native access to your Apple Calendar through system permissions. List, create, update, and delete events without leaving the chat.
Slack
Send messages, browse channels, and stay in the loop — all from inside a chat, persona, or scheduled job.
Cloze
Access your Cloze contacts and activity feed. Look up people, log interactions, and keep your CRM current without switching apps.
Integrations are included with Pro. Connect the services you use and choose which ones your Personas can access.
Choose a different model. Keep the same conversation.
Ask one model to draft and another to review. Bring your files into the discussion and decide who takes the next turn.
Use cloud models through OpenRouter, or connect Codex or Pi on your Mac. Follow the conversation from your phone or browser.
Explore workflows Work from your phone with your MacHow we think about AI agents
Explore our research on context, tools and memory — the ideas behind NanthAI’s agents.
Free to use.
Pro once, use forever.
Start with free cloud chat. Pro is a one-time purchase that adds local runtime and the full toolkit. AI usage is separate: use OpenRouter credit, your Codex subscription, or your configured Pi provider.
£0
A complete multi-model AI chat with basic search, file attachments, folders, and full settings. Not a crippled demo — a real tool.
- Single and multi-model chat (up to 3 models)
- 300+ models — OpenAI, Anthropic, Google, Meta, Mistral, and more
- Basic web search in any conversation
- Image and PDF attachments
- Conversation branching, forking, and regeneration
- Folders, bulk operations, and chat search
- Per-chat model, temperature, and token overrides
- Full model catalog with capability filters
- Native iOS and Android apps
- OpenRouter balance monitoring and low-credit alerts
£4.99
Everything in Free, plus local runtime, Personas, memory, scheduled tasks, integrations, knowledge base, Ideascapes and document creation.
- Local runtime: connect Codex or Pi on your Mac and continue from your phone
- Custom AI Personas for chat, private advice, tasks, and multi-model discussions
- Persistent cross-conversation memory (auto, ask-first, or off)
- Autonomous multi-persona sessions with moderator and consensus detection
- Ideascapes — spatial canvas for non-linear conversations
- Web search at three complexity levels (quick, thorough, comprehensive)
- Research pipeline — plan, search, iterate, and synthesise cited papers
- Scheduled jobs with multi-step pipelines and recurrence
- Knowledge base — upload once, reference from any chat or job
- Document generation (DOCX, XLSX, PPTX, CSV, Markdown)
- Google Drive and Calendar, Manual Gmail, Microsoft 365, Notion, Slack, Cloze, and Apple Calendar integrations
- Parallel subagents — split work across up to 3 focused helpers
- AI skills — attach specialised behaviours to personas
What about model costs?
For cloud chat, connect OpenRouter and pay its listed model rates with no NanthAI markup. Local runtime is included with Pro: Codex uses your existing subscription, while Pi uses your configured provider, subscription or local models.
Provider charges and subscription limits still apply. Models installed through Ollama have no provider token charges. Your OpenRouter balance stays visible in NanthAI, with alerts when credit runs low.
Start with what's free.
Upgrade when you're ready.
Multi-model chat, basic search, file attachments, and folders — all free, no time limit. When you want personas, memory, automations, and the rest, Pro is a one-time unlock.
