Wed 4 Feb 20265 min read

II-Agent V1 is Here

II-Agent V1 is Here

The open-source AI agent built for real work, now out of Beta.

Introduction

Today we're releasing II-Agent V1, the first production-ready version of our open-source intelligent assistant. It delivers state-of-the-art performance, scoring 61.8% on Terminal Bench and 45.1% on SWE Bench Pro.

For months, thousands of builders used II-Agent in beta to ship apps, run research, and automate the work that used to take hours. They found bugs. They pushed limits. They told us what was missing. V1 is the result of that feedback:  a faster, more capable system built for the way people actually work.

II-Agent remains 100% open source under the Apache-2.0 license, and the V1 version will be available for everyone to use in a few weeks. Whether you're a solo developer, a research team, or an enterprise building internal tooling, you can run it, fork it, and extend it with full transparency. No black boxes. No vendor lock-in.

This release isn't a finish line. It's a foundation. We're shipping V1 with the features that matter most today, and we'll keep updating it as the landscape evolves.

What's New in V1

Improved Efficiency 

II-Agent now runs the same prompts in a fraction of the time from the beta version. Delivering faster outputs with more efficient execution, so you can iterate, refine, and ship results dramatically quicker.

Fast Research & Deep Research

Research is the backbone of most knowledge work. V1 ships with two modes built for different needs:

  • Fast Research delivers quick, high-signal answers with key takeaways in seconds; ideal for fact-checking, quick lookups, and staying in flow.
  • Deep Research runs a multi-step investigation across the web, cross-checks sources, and returns a structured brief you can use as a doc, memo, or plan.
  • Interactive Websites: II-Agent can take a deep-research brief (sources, notes, and key claims) and automatically generate an interactive website that explains the findings, turning complex material into a navigable experience with sections, visuals, citations, and embedded Q&A so anyone can explore and understand the research.

Whether you need a fast answer or a comprehensive analysis, II-Agent adapts.

Skills & Custom Skills

We've rebuilt how II-Agent handles repeatable workflows.

Built-in Skills cover common tasks out of the box,  from slide decks to code scaffolding. But the real power is in Custom Skills: connect a GitHub repo, and II-Agent can run your exact processes and reuse them as one-click actions.

This turns II-Agent from a general assistant into a system that knows how your team works.

App Integrations

II-Agent now connects to the tools you already use.

One-click connect to Gmail, Slack, GitHub, Notion, Google Calendar, Discord, Dropbox, Canva, and more. II-Agent can read your emails, check your calendar, pull from your repos, and act across your entire stack, all from a single conversation. More integrations shipping continuously.

Storybook

Generate beautiful, fully illustrated picture books from a single prompt.

Storybook handles the complete creative pipeline: cohesive visual style, consistent characters across pages, and a structured page-by-page narrative. Export and share the finished product. It's creative generation that actually feels finished, not a pile of loose images.

Video & Image Generation

V1 brings a built-in model picker image generation (video coming soon), so you can choose the right engine for the job instead of being locked into a single provider.

Video Model includes VEO 3.1 Premium, with options for quality, duration, multi-scene storytelling, motion realism, audio sync, and fine control.

Image Models include Nano Banana Pro, GPT Image 1.5, and Imagen 4.0 Generate 001 for fast, high-quality visual generation and rapid iteration.

Pick the model. Generate. Iterate. All inside one conversation.

Live Editing

A design-first editor for refining outputs visually.

Website Mode

Slide/Storybook Mode

Change typefaces and fonts, adjust borders, tweak layout and spacing, and dial in the exact styling you want,  all in real time. No more regenerating from scratch when you just need to nudge the kerning.

Everything Else

V1 also includes the full capabilities that made II-Agent a top performer on Terminal Bench 2:

Domain

Capabilities

Full-stack Development Development

Full-stack web builds (React/Next.js, Python, APIs, databases), TypeScript & Tailwind, shadcn/ui, Prisma ORM, sandboxed E2B execution

Documents

PDF extraction & creation, Excel formulas & charts, Word editing, PowerPoint manipulation

Browser

Playwright-powered automation, form filling, screenshot capture, E2E testing

Audio

Transcription and speech synthesis

Slides

Prompt-to-deck creation with live collaborative editing and templates

Plan Mode

Visual project planning before code is written — review the roadmap, then hit Build

Plus full BYOK supports bring your own API keys to control costs and use your preferred providers.

What's Next

V1 is the foundation. Here's what we're building next:

  • More App Integrations: Expanding the Composio-powered library.
  • Agent Development Mobile App: Build and deploy mobile apps faster and more easily than ever. 
  • Custom Personality & Long Term Memory Context: Coming soon. 
  • Expanded Model Support: More frontier models, faster switching.
  • Deeper Agentic Reasoning: Better multi-step planning for complex tasks.
  • Performance Improvements: Faster execution, lower latency.
  • Community-Contributed Skills: A growing library of reusable workflows.

We ship updates continuously. If you're using II-Agent, you're always on the latest version.

Try It Now

Web App: agent.ii.inc

GitHub: github.com/Intelligent-Internet/ii-agent

Community Discord: https://discord.com/invite/intelligentinternet

Join the conversation, share what you're building, and help shape what comes next.

Tutorials 
II-Agent Introduces Skills + Live Editing Mode

How to Create Video, Comics, and Storybooks with Live Editing in II-Agent

II-Agent App Integrations + Universal Connectors

Share this article: