An AI-native, high-performance code editor built on the Zed editor foundation (Rust). Rany IDE runs autonomous background agents that work proactively — analyzing code quality, generating tests, and maintaining documentation — without waiting to be prompted.
The dominant paradigm in AI-assisted coding tools (GitHub Copilot, Cursor, Continue) is reactive: the developer writes, then asks the AI for help. The AI responds, and waits again. This model treats AI as a sophisticated autocomplete.
Rany IDE rejects this model. It is built on the principle that AI agents should be autonomous collaborators — running continuously in the background, proactively contributing to code quality, test coverage, documentation, and security without the developer needing to prompt them. The developer codes; the agents work.
Zed is a multiplayer, GPU-accelerated code editor written entirely in Rust by the creators of Atom and Tree-sitter. It is designed for performance characteristics that no Electron-based editor can match:
Rany IDE builds on all of these properties without modification. The Zed codebase is preserved intact; the autonomous agent system is added as a distinct runtime layer that observes the editor state but does not interfere with its critical path.
The agent system is Rany IDE's primary innovation. Each agent is an independent process that runs with a configurable priority and communicates with the editor via an internal messaging protocol:
Continuously analyzes active files for quality issues, code smells, unnecessary complexity, and deviation from configured coding standards. Suggests refactorings and style improvements proactively.
Monitors for untested functions and paths. Generates unit tests, integration tests, and property-based tests. Detects when existing tests fail to cover recent changes and proposes new test cases.
Generates and updates documentation, API docs, and inline comments. Detects undocumented public APIs and generates contextually accurate JSDoc/Rustdoc/Pydoc comments.
Real-time scanning for vulnerability patterns, dependency CVEs, insecure API usage, and hardcoded secrets. Integrates with Orzatty Labs vulnerability database.
Identifies performance bottlenecks: O(n²) loops, unnecessary allocations, blocking I/O in async contexts, and missing memoization opportunities.
Provides architectural guidance: detects coupling violations, suggests design patterns, identifies circular dependencies, and flags anti-patterns for the active codebase.
Agents communicate with the editor coordinator via an internal async message bus. They observe file open/save events, cursor position changes, and diagnostics — but never block the editor's main thread:
// Agent lifecycle (simplified Rust pseudo-code)
struct CodeQualityAgent {
config: AgentConfig,
sender: MpscSender<AgentMessage>,
}
impl Agent for CodeQualityAgent {
async fn run(&mut self, editor: &EditorObserver) {
loop {
let event = editor.next_event().await;
if let EditorEvent::FileSaved(path) = event {
let analysis = self.analyze(&path).await;
for suggestion in analysis.suggestions {
self.sender.send(AgentMessage::Suggestion {
file: path.clone(),
range: suggestion.range,
message: suggestion.message,
severity: suggestion.severity,
}).await?;
}
}
}
}
}
Each agent is configured independently via a JSON settings block in Rany IDE's configuration file. Two key parameters govern agent behavior:
approval_mode: "auto_safe" — The agent applies safe suggestions (documentation, style fixes) automatically without prompting the developerapproval_mode: "always_ask" — The agent proposes all suggestions in a review panel that the developer accepts or rejects explicitly{
"agents": {
"code_quality": {
"enabled": true,
"approval_mode": "auto_safe", // Apply safe refactors automatically
"priority": 5, // 1 (lowest) to 10 (highest) CPU priority
"trigger": "on_save" // or "continuous" or "on_idle"
},
"testing": {
"enabled": true,
"approval_mode": "always_ask", // Always ask before generating tests
"priority": 4,
"trigger": "on_save",
"test_framework": "jest" // jest | cargo-test | pytest | etc.
},
"documentation": {
"enabled": true,
"approval_mode": "auto_safe",
"priority": 3,
"trigger": "on_idle", // Runs only when editor is idle for 30s
"doc_format": "jsdoc" // jsdoc | rustdoc | pydoc | etc.
}
}
}
Rany IDE inherits Zed's multiplayer collaboration system unchanged. Multiple developers can edit the same file simultaneously with operation-based CRDT conflict resolution — no explicit merging, no overwrite conflicts.
Rany IDE extends this collaboration model to the agent layer: when two developers are collaborating on a session, agent suggestions from any participant's IDE are visible to all collaborators in the shared review panel. This means the Testing Agent's generated test cases are immediately visible to the team — not buried in a single developer's local IDE.
Rany IDE ships with an Orzatty-branded theme as the default visual experience. The color system uses the Orzatty purple/pink brand palette while maintaining readability standards for extended coding sessions:
| Element | Color | Contrast Ratio |
|---|---|---|
| Primary accent (cursor, selection) | #7C4DFF (Purple A200) | — |
| Active tab indicator | #FF4081 (Pink A200) | — |
| Editor background (dark) | #1E1E2E | — |
| Normal code text | #CDD6F4 | 11.7:1 against #1E1E2E ✓ |
| Comments | #6C7086 | 4.6:1 against #1E1E2E ✓ |
| Keywords | #CBA6F7 | 8.2:1 against #1E1E2E ✓ |
# Build from source (requires Rust 1.75+, Node.js 18+)
git clone https://github.com/orzattyholdings/rany-ide.git
cd rany-ide
cargo build --release
# Output binary: ./target/release/rany-ide
# Run tests:
cargo test --workspace
# Property-based tests:
cargo test --workspace --features proptest
Configuration: Stored at platform-standard locations — ~/Library/Application Support/Rany IDE (macOS), $XDG_DATA_HOME/rany-ide (Linux), %LOCALAPPDATA%\Rany IDE (Windows).
| Version | Target | Feature |
|---|---|---|
| 0.1 | Q2 2026 | Public alpha: Code Quality, Testing, and Documentation agents stable |
| 0.2 | Q3 2026 | Security Agent: CVE detection, hardcoded secret scanner, dependency audit |
| 0.3 | Q3 2026 | Performance Agent: complexity analysis, allocation audit, async path check |
| 0.4 | Q4 2026 | Architecture Agent: dependency graph visualization, coupling metrics |
| 1.0 | Q1 2027 | All 6 agents stable. OrzattyAccount SSO for shared agent configurations. OrzattyOS native package. |