AstrolexisAI (@AstrolexisAI) header
AstrolexisAI (@AstrolexisAI) profile picture

AstrolexisAI • Sovereign AI: Kulvex self-hosted platform + KCode local terminal coding agent. 100% GPU, no cloud. mnemo models. Privacy & control first.

Joined March 2026
Photo by AstrolexisAI (@AstrolexisAI)Photo by AstrolexisAI (@AstrolexisAI)Photo by AstrolexisAI (@AstrolexisAI)
Built a code auditor that bundles tree-sitter grammars for 8 languages directly into a single static binary. ~12 MB of grammar wasm, no install dance, runs entirely offline. Open source: github.com/AstrolexisAI/KCod…
1
1
1
2
306
AstrolexisAI retweeted
PyBLOCK × AstroLexis AstroLexis LLM Market is in public beta: open AI models rented by the prompt, paid in XBT sats over Lightning. It runs on PyBLOCK rails: Laguz ᛚ for payments, our Nostr relay for the market. Got a GPU? Sell it for XBT. llm.astrolexis.space
4
7
76
AstroLexis LLM Market is in public beta. Rent open models (Gemma, gpt-oss) straight from someone’s GPU or Mac, prompt by prompt, paid over Lightning in XBT sats. No account. Every seller is benchmarked and audited: the model you pay for is the model that answers. Have a 12 GB+ NVIDIA card or an Apple Silicon Mac? One command and it earns sats. llm.astrolexis.space
2
5
40
The math on cloud LLMs flipped. 2024: cheaper to call someone's API. 2026: cheaper to run on your own hardware after 4-8 months. I wrote up cost, privacy, latency, vendor lock-in — the case for running LLMs on-device in 2026. astrolexis.space/blog/why-on…
134
Official KCode cascade demo thread 👇Two real runs with v2.10.406: • NASA openmct → 13 candidates, 0 false alerts • OWASP Benchmark SQLi → F1 0.829, 100% recall, cascade fully engaged Grok-4-fast + Opus-4.7 ensemble beating single-model Opus by a mile. teapawt.net/curly60e/status/204991…
Two cascade verifier runs on KCode v2.10.406. Same binary, same flags. 🛰️ NASA Mission Control: 13 candidates → 0 false alerts ⚠️ OWASP Benchmark v1.2: 15 candidates → 5 confirmed via ensemble The verifier knows the difference between clean code and real bugs. 🧵
160
Official KCode thread on the SpaceX + Cursor deal 👇 Cursor writes code at lightning speed. KCode makes sure that code has no critical vulnerabilities. In real security analysis, KCode is already a step ahead.
SpaceX just struck a deal with Cursor — up to $60B option to acquire.Cursor is outstanding at writing code fast.But if you actually care about that code being secure… KCode is superior. Straight up
155
After many hours of benchmarking and fighting with LLMs... KCode v2.10.406 is finally out with proper cascade ensemble. This is the one I've been waiting for. Great work team @AstrolexisAI 🔥 @hackerfantastic curious what you think about the Grok + Opus combo.
🚀 KCode v2.10.406 is now available. We just shipped and fully verified end-to-end cascade ensemble (Grok-4-fast → Opus-4.7). Real results on OWASP Benchmark SQLi (1,494 findings): → F1: 0.829 → Precision: 69.8% → Recall: 99.6% → Cost: ~$17 kulvex.ai/kcode/benchmarks
1
3
8
1,053
🚀 KCode v2.10.406 is now available. We just shipped and fully verified end-to-end cascade ensemble (Grok-4-fast → Opus-4.7). Real results on OWASP Benchmark SQLi (1,494 findings): → F1: 0.829 → Precision: 69.8% → Recall: 99.6% → Cost: ~$17 kulvex.ai/kcode/benchmarks
1
1
4
1,182
Download v2.10.406 (Linux, macOS, Windows): → kulvex.ai/kcode Recommended command to try the ensemble: kcode audit . -m grok-4-fast-non-reasoning --fallback-model claude-opus-4-7
1
3
99
This release is another step toward our goal: Best recall in the industry, full control over noise, and reasonable costs. Huge thanks to everyone testing and giving feedback. What do you think of the new cascade? Would you use it in CI? kulvex.ai/kcode/benchmarks
2
29
Replying to @lifeofjer
You can try this. Totally auditable code. teapawt.net/astrolexisai/status/20…
We ship KCode, a security audit tool. Today we audited it against itself and found a bug we'd flag in someone else's code: shell injection in our own PR generator. Fix shipped in v2.10.380. ↓
1
1
3
166
We ship KCode, a security audit tool. Today we audited it against itself and found a bug we'd flag in someone else's code: shell injection in our own PR generator. Fix shipped in v2.10.380. ↓
1
1
3
282
Update size from v2.10.379: 141 KB downloaded 0.8s end-to-end exact SHA match Full binary: 117 MB. Delta = 0.1% of size. Verified E2E against the live manifest at kulvex.ai. No npm, no GitHub Releases — self-hosted auto-updater.
1
1
38
Update: kcode update Install: curl -fsSL kulvex.ai/install.sh | sh For tiny deltas: apt/brew/dnf install bsdiff Repo: github.com/AstrolexisAI/KCod… Release: github.com/AstrolexisAI/KCod…
1
22
The most useful part of the internal audit rounds: found 4 real bugs in my own patterns before they shipped. `new Function(p)` was silently missed (the query only matched the bare-call form). `\eval($p)` PHP namespace-prefixed too. Tree-sitter gives you the AST. You still own the iteration.
Built a code auditor that bundles tree-sitter grammars for 8 languages directly into a single static binary. ~12 MB of grammar wasm, no install dance, runs entirely offline. Open source: github.com/AstrolexisAI/KCod…
1
2
78
Single static binary. No backend. No telemetry. Source code never leaves your machine. The LLM call (when used) only sees the snippet being verified — not the surrounding repo. Apache 2.0.
1
1
33
Looking for feedback on patterns I'm missing — especially from people who've done audit work professionally. Pattern library lives at src/core/audit-engine/ast/ if you want to PR a new sink for your favorite language. github.com/AstrolexisAI/KCod…
1
2
35