100xprompt is an AI coding agent that reads your codebase, edits code across many files, runs commands to verify the change, and keeps going until the task is done. It runs on the most capable frontier models and follows you across terminal, browser, and IDE. This page is the map to everything it can do. Each capability has a dedicated guide - use the cards to dive in. This page covers:

See the capability map

Everything 100xprompt does grows from one loop: understand your project, act on it, verify the result, and remember what it learned. The six branches below are the whole product at a glance - each links to a dedicated guide further down this page.
You stay in control the whole way. 100xprompt asks before it does anything sensitive, and every change is yours to review, refine, or roll back.

Understand the core loop

Every capability plugs into the same cycle. You describe an outcome. 100xprompt figures out the steps, does them, checks its work, and carries the context forward - so the next task starts smarter than the last.

Build

100xprompt writes and changes real code, understands the project it works in, and runs the commands to prove the change works. This is the core.

Edit code across many files

Make coordinated changes spanning dozens of files - refactors, features, and fixes - with edits applied precisely and reviewable diff by diff.

Understand large codebases

Search, navigate, and trace relationships across an entire repository so answers and changes are grounded in how your code actually works.

Run commands & tests

Execute builds, test suites, linters, and scripts - then read the output and iterate until the result is green.

Multi-step agentic tasks & subagents

Break big goals into steps, and spin up focused subagents to explore or work in parallel without losing the main thread.

Think

100xprompt reasons before it acts. It picks the right model for the job, remembers your project across sessions, and keeps the most relevant context in view on long tasks.

Choose frontier models

Switch between the most capable frontier models per task - reach for deep reasoning on hard problems and fast models for quick edits.

Persistent memory & context

Carry facts, decisions, and preferences across sessions, and keep long tasks coherent with automatic context management.
Models and memory work together: pair a high-reasoning model with rich project memory for your hardest changes, then drop to a faster model for routine follow-ups.

Extend

Teach 100xprompt your stack and connect it to your tools. The same agent gains new abilities as you add Tools, Skills, Plugins, and MCP connections.

Tools

The built-in actions 100xprompt uses to read, edit, search, and run - the foundation every task is built on.

Skills

Package reusable expertise and procedures so 100xprompt handles specialized work your way, every time.

Plugins

Install bundles of skills, commands, and connections to extend what 100xprompt can do in one step.

MCP connections

Connect databases, SaaS tools, and internal services so 100xprompt can act on the systems your work depends on.

Automate

Move work off your plate. Run tasks in the background while you stay in flow, put routine jobs on a schedule, and turn repeatable procedures into workflows. Reviewing pull requests by hand eats your day. Hand the routine passes to 100xprompt and it runs them on a schedule, reporting back only what needs your eyes.

Background & scheduled tasks

Kick off work that runs on its own - asynchronously or on a recurring schedule - and review the results when they’re ready.

Repeatable workflows

Capture multi-step procedures once and run them on demand, so the same job is done the same way every time.

Work anywhere

One agent, three surfaces. Start a task in your terminal, hand it off in the web app, and keep the same context in your editor. The same session follows you wherever you go.

Terminal

A fast, first-class terminal UI for driving 100xprompt without leaving your shell.

Web

A full web app to run sessions, review changes, and collaborate from anywhere.

IDE

Bring 100xprompt into your editor so changes land right where you’re already working.

Share and collaborate

Bring your team along. Share sessions and results, and roll out 100xprompt across an organization with the controls enterprises need.

Collaboration & Enterprise

Share sessions, standardize how your team works, and roll out 100xprompt with the security and governance an organization requires.

Capabilities at a glance

CapabilityWhat it doesLearn more
Code editingCoordinated, reviewable changes across many filesEditing code
Codebase understandingSearch, navigation, and relationship tracingUnderstanding codebases
Running commandsBuilds, tests, linters, and scripts with iterationRunning commands
Agents & subagentsMulti-step tasks and parallel focused agentsAgents
ModelsChoose the right frontier model per taskModels
Memory & contextPersistent memory and long-task context handlingMemory & context
ToolsBuilt-in read, edit, search, and run actionsTools
SkillsPackaged, reusable expertiseSkills
PluginsInstallable bundles of extensionsPlugins
MCPConnections to external systemsMCP
AutomationBackground, scheduled, and repeatable workWorkflows
TerminalFirst-class terminal experienceTerminal
WebFull-featured web appWeb
IDEEditor integrationIDE
EnterpriseSharing, security, and governanceEnterprise

Choose where to start

Not sure which capability you need first? Route by what you want to do:
I want to…Best option
See a real change land, end to endEdit code: watch a multi-file change from prompt to verified diff
Understand an unfamiliar codebaseUnderstand codebases: ask questions grounded in your code
Make it fluent in my stackSkills: package your team’s procedures and conventions
Connect my own tools and dataMCP connections: reach databases, SaaS, and internal services
Take work off my plateWorkflows: run tasks in the background or on a schedule

See it edit code

The fastest way to feel what 100xprompt does - a real change, end to end.

Extend it to your stack

Add Skills, Plugins, and MCP connections to make it fluent in your tools.

Quickstart

Go from an empty terminal to your first shipped change in minutes.

CLI Reference

Every command and flag, grouped by what you reach for it to do.

Models

Choose the right frontier model for each task.