# Akon Labs: GitNexus gives coding agents a resolved graph of the codebase

Canonical: https://mudpie.ai/companies/akon-labs/
Breadcrumb: [Home](https://mudpie.ai/) / [Companies](https://mudpie.ai/companies/) / [Akon Labs: GitNexus gives coding agents a resolved graph of the codebase](https://mudpie.ai/companies/akon-labs/)
Author: Ali Abouelatta (https://mudpie.ai/authors/ali-abouelatta/)
Published: 2026-09-19
Updated: 2026-09-19
Research type: Company profile
Method: Company and accelerator sources checked 2026-09-19. Product claims are attributed to their sources; this is research, not a hands-on product trial.

Akon Labs builds GitNexus, a code knowledge graph that gives coding agents resolved dependencies, call chains, architecture clusters, and blast-radius analysis. The buyer decision is whether exact structural context will improve agent work enough to justify a graph layer—especially in a large, polyglot, regulated, or air-gapped codebase.

## What it does

GitNexus indexes repositories with compiler-style parsing, resolves imports and call paths, clusters the architecture, maps a diff to affected symbols, and exposes the graph through MCP tools and editor integrations. Akon Labs offers the open-source engine locally, a managed Enterprise version with multi-repo graphs and PR review, and self-hosted deployment for teams that cannot send code outside their network ([Akon Labs homepage](https://www.akonlabs.com/); [Akon Labs pricing](https://www.akonlabs.com/pricing)).

| Fact | What the public sources say |
| --- | --- |
| Buyer | Engineering and platform teams using coding agents across large repositories |
| Core output | Dependency graph, execution flows, architecture clusters, impact analysis, and MCP context |
| Deployment | Local open source, managed SaaS, or self-hosted enterprise |
| Public pricing | Open source free; managed Enterprise listed at $29/seat; self-hosted custom |
| Founder context | YC lists Subham Kundu; the launch also names Abhigyan Patwari on the GitNexus team |

## Why it fits

The product's useful distinction is resolved structure. Grep and embeddings can find similar text; GitNexus aims to answer which callers, callees, imports, and downstream services are actually connected. That is the kind of context an agent needs before changing a shared function or proposing a PR across several repositories. The local and air-gapped options also make the architecture relevant to teams with strict code-boundary requirements.

The homepage and pricing page point to a DeepSWE benchmark where the graph-assisted setup solved 68.4% of tasks versus 37.0% for the comparison setup. That is a company-published benchmark result, not proof that every production repository will see the same lift ([Akon Labs benchmark page](https://www.akonlabs.com/benchmarks)). A buyer should test dynamic imports, generated code, polyglot services, stale branches, permissions, and the cost of keeping the graph current.

Kundu's public background includes CTO work at Cignara, AI observability at HTCD, and an earlier LLM chatbot at Caravel Labs; the launch describes Kundu and Patwari as having spent years on knowledge graphs and open-source systems ([YC company profile](https://www.ycombinator.com/companies/akon-labs); [YC launch](https://www.ycombinator.com/launches/Smp-akon-labs-we-re-building-the-most-token-efficient-coding-agent-in-the-world)).

Short version: start with GitNexus when agents repeatedly miss cross-file impact. The decisive proof is a representative monorepo and a change that currently makes your agent guess.

## Sources checked — 2026-09-19

- [YC company profile](https://www.ycombinator.com/companies/akon-labs)
- [Akon Labs homepage](https://www.akonlabs.com/)
- [Akon Labs pricing](https://www.akonlabs.com/pricing)
- [Akon Labs benchmark page](https://www.akonlabs.com/benchmarks)
- [Akon Labs YC launch](https://www.ycombinator.com/launches/Smp-akon-labs-we-re-building-the-most-token-efficient-coding-agent-in-the-world)

## Cohort context

Akon Labs is listed in Summer 2026. In our 2026-09-18 directory snapshot, 119 of 232 listed companies in that cohort have YC’s primary industry label B2B (51.3%). This is a current-directory comparison, not an original intake count or a performance ranking. [Nine-cohort dataset](https://mudpie.ai/research/yc-cohorts-2026-09-19.json).

## Public website snapshot

Observed 2026-09-19T16:18:32.629Z in raw homepage HTML. This records visible metadata and advertised links, not agent execution or product quality.

| Signal | Homepage observation |
| --- | --- |
| Product description metadata | Observed |
| Canonical link | Observed |
| H1 or H2 heading | Observed |
| Typed structured data | Not observed in this response |
| Docs/developer link | Not observed in this response |
| Pricing link | Observed |
| llms.txt link | Not observed in this response |
| Markdown alternate | Not observed in this response |

[Public observations](https://mudpie.ai/research/yc-homepage-links-2026-09-19.json) · [Collection method](https://mudpie.ai/research/yc-homepage-methods/README.md). Missing links here do not establish that a capability or file is absent elsewhere.


## Author disclosure

I cofound Lazyweb and publish Mudpie. This is an owner-written publication, not an independent testing organization. Research notes distinguish observations, sourced reporting and editorial judgment.