I'm a student building a code intelligence platform. I ran it head-to-head against two established competitors. Here's what actually happened.
Quick intro, since this is my first post here: I'm Arihant, a student in India, building Aletheore alone, alongside two degrees. It's been public for about six weeks. This post is what it is, why it's different, and two real (not cherry-picked) head-to-heads against established competitors in the space. What it does Aletheore is a code intelligence platform: it parses a repository into a real…
Arihant, a student from India, has been building a code intelligence platform called Aletheore alone alongside two degrees. The platform parses a repository into a dependency graph and grounds everything downstream in that graph. This approach enables features like security scanning and PR review, as well as generated architecture documentation, dead-code detection, and semantic code search.
Aletheore provides these features through a CLI, MCP server, GitHub App that comments on PRs, and a live architecture dashboard. The platform offers a free tier, making it accessible to users without needing to install it.
In a head-to-head comparison with two established competitors, RepoWise and Graphify, Aletheore demonstrated both superiority and areas where it fell short. In the first test, Aletheore outperformed RepoWise in code locating tasks, securing all 5 wins and no losses. However, Aletheore struggled with top-5 results in the Jekyll corpus, where RepoWise achieved a higher accuracy of 66.7%, compared to Aletheore's 46.7%.
In terms of natural language vocabulary questions, RepoWise closed some of the gap as its wiki pages name symbols directly, but still falls short of Aletheore's performance.
In the second head-to-head against Graphify, Aletheore and Graphify demonstrated similar performance on a coverage metric measured against the frappe/erpnext codebase. However, Aletheore achieved a slight edge in token cost, using 36% fewer tokens to provide the same answers. The comparison also highlighted the importance of setup time, with Graphify taking approximately a minute to build its whole graph, while Aletheore took around 23 seconds.
After identifying and addressing an inefficient algorithm in Aletheore's reference check, its setup time was reduced from 236 seconds to 53 seconds, representing a 4.4x improvement.
Aletheore's live-docs feature offers a unique advantage in generating AI-written architecture pages for repositories, but it does not cover every file by default. However, a deterministic fallback method that reads the scanner's own module record makes Aletheore's feature reliable, covering all files with no marginal cost per file. The platform's main focus is not the generated documentation, but the ability to provide reliable PR reviews without depending on a wiki page existing for every file.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
