We built Jolt to solve a fundamental problem with AI coding tools - they struggle with larger, real-world codebases and cannot accurately determine the context files for your prompt.
Jolt's public beta is live today, and we'd love your feedback. You can use Jolt on the web or in any VSCode-based IDE. Our free tier will let you work on codebases up to ~100K lines.
We developed a novel way for Jolt to understand large codebases and automatically determine context files. Our approach scales to multi-million line codebases - the largest codebase using us in production is over 8M lines.
Jolt is optimized for codebase understanding rather than response speed. That said, using Jolt is significantly faster than other tools when you include the time spent selecting context files.
Manually selecting context files is a non-starter when you're working on larger codebases or unfamiliar code. It's a broken product experience. Most AI coding tools rely on some flavor of vector-embedding RAG to determine the files related to your prompt. The reality is that vector-embedding search is not effective on code, and there is a sharp drop in efficacy as codebase size increases.
Here's what folks are using Jolt for: - Writing code, tests, and refactoring - Onboarding developers - Brainstorming a feature's implementation - Asking questions about OSS repos - Writing documentation, including mermaid charts - Contributing across the stack
Thanks for reading, and let us know what you think.