KodaCode rolls out agent control via Telegram and voice input
KodaCode has added two key features for remote agent control. Developers can now send commands from Telegram, check execution status, and get results directly on their phone — without needing to open an IDE. Voice input lets them speak commands in natural language, without worrying about exact phrasing. The two modes complement each other: Telegram for remote access when the IDE isn't available, and voice for quickly dictating complex tasks. The KodaCode team also shipped a new autocomplete mode and fixed tool-parsing bugs that were causing responses to render incorrectly. Version 0.9.0 is ready to use. • Remote control via Telegram (available in VS Code, planned for CLI and JetBrains) • Voice command input (in VS Code, extension in development) • New autocomplete mode and parsing fixes (version 0.9.0)
AI-processed from Habr AI; edited by Hamidun News
KodaCode has shipped two of its most anticipated features for controlling its IDE agent: remote control via Telegram and voice command input. It's a solution for developers who often work away from their desk or want to speed up how they interact with the agent. Both capabilities have already appeared in leading international tools like Codex and Claude Code — KodaCode is now adding them to its own platform, giving developers the flexibility to work with the agent from anywhere and however is most convenient.
Remote control via Telegram
Before this release, the KodaCode agent was tightly tied to the IDE or terminal. As long as the developer was at the computer, they could edit code, run commands, and work on the project. But step away from the desk, and access to the agent disappeared. An idea strikes on the road, there's an urgent need to run a script on a remote server, check a build status — but the IDE is closed on the laptop, and solving the problem on the spot was impossible.
Now KodaCode supports interacting with the IDE agent through a Telegram bot. Developers can send commands from the messenger and receive results there too: kick off a task, check execution status, get the result — all straight from the phone. The team didn't choose Telegram by accident: the messenger is actively integrating with AI tools, and Telegram itself has long been a working tool for most developers.
Setup instructions for remote access are built into the plugin itself (Settings → Enable Telegram remote → Telegram). Remote control is already available in VS Code; it will soon be added to the CLI and JetBrains IDEs. The team is aware of the current complexities around Telegram and is already discussing which alternative remote access channels to implement first.
Voice command input
As the pace of development accelerates, speaking to the agent is becoming a natural choice — not just when your hands are busy or typing isn't an option, but simply because it's faster. The KodaCode agent doesn't require precise phrasing or complete sentences: you can talk the way you normally would, with pauses, filler sounds like "uh," without finishing a sentence — the agent figures out what needs to be done.
The interface is as simple as it gets: just tap the microphone icon and start talking. Voice input is available in VS Code; it will soon be added to the CLI and JetBrains IDEs.
Other improvements in version 0.9.0
Beyond the two headline features, the 0.9.0 release of KodaCode includes the following changes:
• A new autocomplete mode that predicts the next edit in the context of changes already made, rather than simply completing at the current cursor position • A fix for tool-call parsing and argument streaming that was causing the agent's responses to render incorrectly • Improved chat history recovery and reloading of the last session • Performance optimizations
The pace of development is clearly picking up: in the weeks leading up to this release, Telegram has been actively rolling out new capabilities, integrating deeper support for AI agents. KodaCode is tracking these trends and adapting accordingly. Detailed documentation on KodaCode, including articles on the JetBrains IDE, VS Code, and CLI versions, is available on Habr.
- Oct 27, 2026Fudan and Huawei Xiaoyi Teach an AI Agent to Write Operators for Ascend NPUs: 95.5% Success Rate
- Oct 27, 2026EmbodiedLab: robots tested on experiments, and the best one solves less than half
- Oct 27, 2026Skill∞: an agent improves its own skills and its own meta-skill, 60.7% on SkillsBench
- Oct 27, 2026Work Agents: 16 Researchers Build a Unified Coordinate System for AI Agents
Key Questions Answered
What are the key details on Remote control via Telegram?
Before this release, the KodaCode agent was tightly tied to the IDE or terminal. As long as the developer was at the computer, they could edit code, run commands, and work on the project. But step away from the desk, and access to the agent disappeared. An idea strikes on the road, there's an urgent need to run a script on a remote server, check a build status — but the IDE is closed on the laptop, and solving the problem on the spot was impossible. Now KodaCode supports interacting with the IDE agent through a Telegram bot. Developers can send commands from the messenger a
What are the key details on Voice command input?
As the pace of development accelerates, speaking to the agent is becoming a natural choice — not just when your hands are busy or typing isn't an option, but simply because it's faster. The KodaCode agent doesn't require precise phrasing or complete sentences: you can talk the way you normally would, with pauses, filler sounds like "uh," without finishing a sentence — the agent figures out what needs to be done. The interface is as simple as it gets: just tap the microphone icon and start talking. Voice input is available in VS Code; it will soon be added to the CLI and Jet
What are the key details on Other improvements in version 0.9.0?
Beyond the two headline features, the 0.9.0 release of KodaCode includes the following changes: A new autocomplete mode that predicts the next edit in the context of changes already made, rather than simply completing at the current cursor position A fix for tool-call parsing and argument streaming that was causing the agent's responses to render incorrectly Improved chat history recovery and reloading of the last session Performance optimizations The pace of development is clearly picking up: in the weeks leading up to this release, Telegram has been actively rolling out n
Build & Ship Production AI Agents Without Boilerplate Code
Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.