
TL;DR
Cursor v3.11 introduces Side Chats for parallel agent conversations, Conversation Search across past sessions, and Cloud Agent Hooks for self-correcting loops. A practical guide to the new features released July 10, 2026.
Cursor v3.11 shipped July 10, 2026 with one feature developers have been asking for since Composer became the default experience: parallel conversations. You can now spin up side chats that explore tangents without interrupting your main agent session.
Last updated: July 11, 2026
| Resource | Link |
|---|---|
| Cursor Changelog | cursor.com/changelog |
| Cursor Documentation | docs.cursor.com |
| Cursor Blog | cursor.com/blog |
| Cursor Pricing | cursor.com/pricing |
| Cursor Forum | forum.cursor.com |
Side Chats let you open parallel agent conversations that run alongside your main chat. Each side chat is a full agent session - it can read files, run commands, and make edits - but it stays separate from your primary thread.
Three ways to open a side chat:
/side or /btw in your main chatEach side chat inherits context from the main chat at the moment you spawn it. You can then take it in whatever direction you need - explore an alternative approach, research a library, debug a specific function - without polluting your main conversation.
Before v3.11, exploring tangents meant either:
Side Chats solve the context problem. The main thread keeps running. You can fire off a quick question, get an answer, and either discard the side chat or pull its findings back into the main conversation with an @mention.
The workflow pattern Cursor is enabling:
Main: "Build the user authentication system"
└─ Side 1: "@btw what's the best JWT library for Deno?"
└─ Side 2: "@btw can you check if our current session middleware handles refresh tokens?"
Main: "Thanks @side-1, let's use jose. And @side-2 confirmed we need refresh token handling."
Each side chat is durable. You can close it, come back later, and continue the conversation. The @mention syntax pulls context from side chats into your main thread.
The second major feature: searchable agent history. Cursor now builds a local index of your past conversations.
Global search (Cmd+K in Agents Window): Search across all your past agent chats. This goes beyond filenames and PR numbers - you can search for concepts, error messages, or approaches you discussed weeks ago.
In-chat search (Cmd+F): Jump between search matches within a single conversation. The counter shows your position in results.
This addresses a real pain point. With heavy Composer usage, you accumulate hundreds of agent conversations. Finding "that chat where I figured out the Redis connection pooling issue" used to mean scrolling through history or relying on memory. Now you can search for "Redis pool" and find it.
Newsletter
Get the weekly deep dive
Tutorials on Claude Code, AI agents, and dev tools, delivered free every week.
From the archive
Jul 11, 2026 • 6 min read
Jul 11, 2026 • 6 min read
Jul 10, 2026 • 6 min read
Jul 10, 2026 • 8 min read
For teams building on top of Cursor's agent infrastructure, v3.11 adds programmable hooks:
| Hook | When It Fires |
|---|---|
beforeSubmitPrompt | Before your prompt is sent to the model |
afterAgentResponse | After the agent generates a response |
afterAgentThought | After each reasoning step (for extended thinking models) |
stop | When the agent stops for any reason |
subagentStart | When a subagent spawns |
These enable patterns like:
Hooks are configured in your project's .cursor/hooks.json and run in Cursor's cloud execution environment.
Smaller but useful: the project and repo selection UI got an overhaul.
Consolidated workflows: Creating projects, connecting GitHub/GitLab/Azure DevOps, and switching repos all happen in the picker. No more bouncing between settings screens.
Scoped search: Search is now contextual - "This Computer," "Cloud," or specific remote machines. The old global search box sometimes surfaced confusing results mixing local and remote repos.
Branch picker defaults: Opens to your recent branches instead of alphabetical. Find "no repo" by typing "none" or "no repo."
Here's how to get the most from Side Chats:
You're in the middle of implementing a feature when you hit a library question:
Main: "Implement rate limiting on the /api/process endpoint"
Agent: [working on implementation]
You: /side what rate limiting libraries work with Hono?
The side chat researches options while your main chat continues the implementation. When ready, @mention the findings back.
Something's broken and you want to investigate without losing your main context:
Main: "The tests are passing but production throws a null reference"
You: /btw can you check the error handling in services/processor.ts?
The side chat digs into the specific file. If it turns out to be a red herring, close the side chat. If it finds the bug, pull the fix into main.
You're not sure about the agent's suggested approach:
Main: "Let's use a recursive approach for the tree traversal"
You: /side what would the iterative version look like?
Compare both approaches without abandoning either conversation. Useful when you want to evaluate trade-offs before committing.
Preparing to hand off work to a colleague:
Main: [your ongoing implementation work]
You: /side summarize what we've built so far and what's left
Generate handoff documentation without interrupting your flow.
| Action | Mac | Windows |
|---|---|---|
| New Side Chat | Cmd+Shift+N | Ctrl+Shift+N |
| Global Conversation Search | Cmd+K (in Agents Window) | Ctrl+K |
| In-Chat Search | Cmd+F | Ctrl+F |
| Next Search Match | Cmd+G | Ctrl+G |
| Previous Search Match | Cmd+Shift+G | Ctrl+Shift+G |
No pricing changes with v3.11. Side Chats consume requests from your existing plan allocation - Pro, Teams Standard, or Teams Premium.
Side chats run as full agent sessions, so they do count against your usage. If you're on the free tier, heavy side chat usage will hit limits faster. For paid plans, the additional flexibility is worth it.
VS Code 1.128 shipped multi-chat Claude sessions a few days earlier (July 8, 2026). How do they compare?
| Feature | Cursor v3.11 | VS Code 1.128 |
|---|---|---|
| Parallel conversations | Yes (Side Chats) | Yes (Multi-Chat) |
| Spawn with command | /side, /btw | Fork from turn |
| Context inheritance | At spawn time | Full history fork |
| @mention back | Yes | Yes |
| Conversation search | Yes (Cmd+K) | Yes |
| Cloud hooks | Yes | No |
| Model selection | Per-chat | Per-chat |
The key difference: Cursor's Side Chats are designed for quick tangents - they inherit context at spawn but diverge immediately. VS Code's fork model preserves full history in both branches, which is heavier but useful for different scenarios.
Yes. Each side chat is a full agent session and consumes requests from your plan.
Not directly. You can create a new main chat and @mention the side chat content, or copy-paste key findings.
Yes. Side chats are durable and stored locally. Reopen them from the Agents Window.
Yes. Each side chat can use a different model than your main chat.
Yes. Hooks fire for all agent sessions including side chats.
Yes. Conversation Search is available on all plans.
Yes. Global search (Cmd+K in Agents Window) indexes all conversations including side chats.
No documented limit. Practical limits depend on your system resources and usage patterns.
Read next
VS Code 1.128 shipped today with multi-chat support for Claude agent sessions. Run parallel conversations in one workspace, fork turns, compare approaches, and monitor subagents. Complete setup and workflow guide.
7 min readCursor and Devin Desktop have converged on similar pricing but diverged hard on philosophy. Here is what actually matters when picking one for your team in 2026.
8 min readCursor is editor-first. Codex is terminal, cloud, and PR-first. Here is when to use each for TypeScript projects.
5 min readTechnical content at the intersection of AI and development. Building with AI agents, Claude Code, and modern dev tools - then showing you exactly how it works.
Codeium's AI-native IDE. Cascade agent mode handles multi-file edits autonomously. Free tier with generous limits. Stron...
View ToolMac app for running parallel Claude Code, Codex, and Cursor agents in isolated workspaces. Watch every agent work at onc...
View ToolA hosted infinite canvas your headless AI agents drive over MCP. Any MCP-speaking agent - Claude Code, Codex, Cursor, or...
View ToolAI-native code editor forked from VS Code. Composer mode rewrites multiple files at once. Tab autocomplete predicts your...
View ToolEvery coding agent in one window. Stop alt-tabbing between Claude, Codex, and Cursor.
View AppReplay every MCP tool call to find why your agent went sideways.
View AppTurn a one-liner into a working Claude Code skill. From idea to installed in a minute.
View AppA concrete step-by-step guide to moving your development workflow from Cursor to Claude Code - settings, rules, keybindings, and the habits that transfer.
Getting StartedConfigure Claude Code for maximum productivity -- CLAUDE.md, sub-agents, MCP servers, and autonomous workflows.
AI AgentsStep-by-step guide to building an MCP server in TypeScript - from project setup to tool definitions, resource handling, testing, and deployment.
AI Agents
In this video, discover how to build your customized voice AI agents using TEN Agent, an open-source conversational AI platform. Learn to integrate top speech-to-text models, large language...

Learn The Fundamentals Of Becoming An AI Engineer On Scrimba; https://v2.scrimba.com/the-ai-engineer-path-c02v?via=developersdigest Exploring Cursor's New Agentic Capabilities: A Hands-On...

Learn The Fundamentals Of Becoming An AI Engineer On Scrimba; https://v2.scrimba.com/the-ai-engineer-path-c02v?via=developersdigest Exploring Codeium's New Windsurf Editor: The Future of AI-assist...

xAI launched Grok 4.5, trained on trillions of Cursor interaction tokens. At $2/M input pricing, it undercuts Claude and...

Cursor just shipped Composer 2 - a major upgrade to their AI coding assistant. Here is what changed and why it matters.

A new essay argues that letting AI generate sloppy code creates a downward spiral where future AI absorbs those bad patt...

A Haskell Foundation board member explains why Scarf moved to Python after 7 years in production. The culprit: LLM-drive...

A viral post argues AI works better on standardized codebases, making rewrites economically sensible. HN pushes back wit...

The Bun runtime completed an AI-assisted rewrite from Zig to Rust, fixing memory safety issues and improving performance...

New tutorials, open-source projects, and deep dives on coding agents - delivered weekly.