Jcode

High performance coding agents

2.2Score#642 of 999 agents#28 of 40 in Software engineer

Why it ranks #642

Jcode, an AI software engineer from YC S26, ranks #642 of 999 AI agents and #28 of 40 in Software engineer, with a score of 2.2 of 10. Data as of 3 Oct 2026.

The score adds four parts, each 0 to 10, by weight. Every point comes from public evidence. How scoring works ·

Proof 0 × 30% 0.0
Scale 0 × 30% 0.0
Momentum 4 × 25% 1.0
Autonomy 8 × 15% 1.2
= 2.2

Proof
Do customers use it? 30% of the score
0
No public claims of customers or revenue yet.
Not found: revenue or growth (+4), named customers (+3), a customer result (+2), a customer count (+2), work done at scale (+2)
Scale
Is there a real company behind it? 30% of the score
0
1 person.
Not found: funding round in its news (+3)
Momentum
Is it shipping and growing now? 25% of the score
4
Shows a recent launch and a new batch.
- +3
  Launched in the last 6 monthsJcode: High-performance coding agentsLaunch YC post · 10 Aug 2026
- +1
  New: YC Summer 2026
Not found: open roles on YC (+2), new traction claim in the last 6 months (+2), funding news in the last year (+2)
Autonomy
How much of the job does it do on its own? 15% of the score
8
Autonomous worker: completes whole tasks; people handle exceptions.
- +8
  Level 3 of 4: Autonomous workerCompletes whole tasks; people handle exceptions.
What the model read
- Acts in other systems: 73% likely
- Finishes whole tasks: 54% likely
- Calls itself an agent: 89% likely
Level 3: Completes whole tasks; people handle exceptions.

About Jcode

Jcode is a high-performance coding-agent harness that makes it possible for developers to scale up multi-agent/swarm workflows. It's built from the ground up with a different architecture to solve the memory bottleneck (it's 20x more memory efficient than Claude Code) that has kept agent swarms from going mainstream. It has 20K GitHub stars, and throughputs $1mil per month of inference spend