Blog & news

Product updates, launch notes, and field reports from the TaskForceAI team.

Research

Frontier Model Benchmarks

Current Artificial Analysis comparisons for Sentinel, GPT-5.5, Gemini 3.1 Pro Preview, Claude Opus 4.8, and Grok 4.3.

View benchmarks
Agents4 min read

Computer use comes to your machine and the cloud

TaskForceAI agents can now operate a real computer — locally through the desktop app or in a cloud Linux desktop — with a live theater view of every action.

Read more
Agents4 min read

A real coding workspace, local and remote

The desktop app and TUI now run a local coding agent with a file tree, diff preview, and terminal — and can drive a remote machine over SSH.

Read more
Memory3 min read

Memory you can see, check, and trust

TaskForceAI added desktop Screen Memory controls and made stored memory reviewable, with clear provenance for where each remembered fact came from.

Read more
Tools3 min read

Sharper web search and sandboxed code execution

Agents get deeper web search with visible sources and a sandboxed code execution tool, so they can both gather current information and actually run code.

Read more
Continuity3 min read

Pair mobile with desktop and keep working

You can now pair the mobile app with desktop, follow a workspace from either device, and pick up where you left off through a shared continuity sidebar.

Read more
Launch5 min read

Introducing TaskForceAI

A first look at the TaskForceAI multi-agent orchestration platform, how it works, and what shipped in the public developer preview.

Read more