The Deska blog
Blog
Product notes, release highlights and how we build Deska: the one infinite canvas workspace for your editor, terminal, browser, notes and AI agents.
· 10 min read
Agent Memory Systems: What Persists Between Sessions
Explore how agent memory systems handle persistence across sessions, from short-term context to long-term architectural strategies for AI developers.
· 11 min read
Memory Files: Persistent Agent Knowledge That Works
Learn how to use memory files to provide persistent agent knowledge and improve LLM coding performance across sessions and complex codebases.
· 11 min read
Mapping a Feature to Every File It Touches
Learn how to perform mapping a feature to every file it touches to improve code navigation, reduce regressions, and master complex software architectures.
· 11 min read
An Agent Loop From Scratch in 200 Lines
Learn how to build an agent loop from scratch using Python to understand the core architecture of modern AI coding assistants and autonomous developer tools.
· 10 min read
Making Logs Consistent Enough to Query
Learn how to implement structural logging strategies to make logs consistent enough to query when building and debugging complex AI coding agent systems.
· 11 min read
Auditing Loading and Empty States With an Agent
Learn how to conduct a thorough loading and empty states audit using AI agents to ensure UI completeness and a better user experience in your web applications.
· 12 min read
Writing k6 Scenarios With an Agent
Learn how to accelerate your performance testing by writing k6 scenarios with an agent to simulate realistic user traffic and complex load patterns.
· 12 min read
Where Agents Struggle: Product Judgment
Explore why AI agents struggle with product judgment and how developers can bridge the gap using workspace tools like Deska to maintain quality.
· 10 min read
Where Agents Struggle: Performance Tuning
Explore why AI agents struggle with performance tuning and how a professional workspace helps developers manage complex profiling and optimization tasks.
· 11 min read
Where Agents Struggle: Undocumented Legacy Context
Discover why AI agents struggle with undocumented legacy context and how a local-first workspace like Deska helps developers bridge the gap.
· 12 min read
Where Agents Struggle: Concurrency and State
Explore why AI agents struggle with concurrency and state, and how local-first tools like Deska provide the visibility needed to debug these complex issues.
· 11 min read
Working Through Lighthouse Complaints With an Agent
Learn how to optimize web performance by working through Lighthouse complaints with an agent while maintaining a local-first development workflow.
· 10 min read
License Headers Across a Repo in One Pass
Learn how to automate license headers across a repo in one pass using scripts and AI agents to ensure legal hygiene and open source compliance throughout your code.
· 10 min read
Do Agents Learn From Your Corrections? Sort Of
Understand how AI agents learn from your corrections during development and how context windows impact long term behavioral changes.
· 12 min read
Landing Copy Variants Worth Testing
Explore landing copy variants worth testing for developer tools and AI agents to improve conversion through clear value propositions and technical clarity.
· 10 min read
JS to TS, File by File: An Agent Migration Cadence
Learn a file by file JS to TS migration strategy using AI coding agents. A technical guide on incremental typing, strictness, and workspace orchestration.
· 10 min read
jQuery to Modern JS: The Agent Migration Path
Learn how to use AI coding agents to navigate the jQuery to modern JS transition while maintaining legacy stability and performance.
· 10 min read
Issue Templates That Improve Bug Reports
Learn how to build issue templates that improve bug reports for humans and AI agents, reducing friction in open source and internal development workflows.
· 11 min read
What Happens When You Interrupt Each Agent
A technical guide on agent interrupt behavior across Claude Code, Codex CLI, and OpenCode, and how to manage state recovery in your developer workspace.
· 12 min read
Input Validation Sweeps: Closing the Gaps by Agent
Learn how to perform a comprehensive input validation sweep using AI agents to find security gaps in your code and fix them within an infinite canvas.