glm-5.2
7d7b99c04d
Documents the architectural direction for a PyTorch-shaped tensor computation library built on Rust + wgpu, where QuickJS is a thin API/composition layer and Rust owns memory, dispatch, and WGSL codegen. Derived from webgpu-torch as the reference design (op_spec → opgen → WGSL shader pipeline) but not a port of its code — webgpu-torch is the reference, alknet-tensor is the production architecture. Key decisions: JS holds handles (BufferId), Rust owns wgpu::Buffers; ~4-5 high-level Rust ops (create_tensor/dispatch_kernel/register_kernel/read/write) not ~20 low-level GPU API calls; WgslGenerator as a third handlebars backend in typebox-rs codegen alongside RustGenerator and TypeScriptGenerator; tensor ops as OperationSpecs on the registry (network-callable over irpc, verified protocol-compatible on quickjs by POC 2). Documents the downstream problems this solves as a side effect: distributed compute over irpc, LLM-authored model code (toolEnv pattern), edge/embedded tensor compute, the compositing problem sidestepped (compute has no surface), and cross-platform by construction (wgpu's many backends).
Alknet
Status: Pre-alpha — This project is undergoing a major architectural pivot to an ALPN-as-service model. The previous implementation has been archived and a greenfield rebuild is in progress.
A self-hostable networking toolkit built on QUIC+TLS with ALPN-based protocol dispatch. Each protocol handler (SSH, SFTP, Git, HTTP, DNS, messaging, call protocol) registers an ALPN string on a shared endpoint. The ALPN negotiation during the TLS/QUIC handshake routes connections to the correct handler before any application bytes are read.
Core Insight
A service IS an ALPN. One endpoint, one port, many protocols — dispatched by the TLS handshake, not by application-level peeking or separate listeners.
Crates
| Crate | Status | Description |
|---|---|---|
alknet-vault |
stable | Local key vault: BIP39/SLIP-0010/AES-GCM key derivation and encryption |
alknet-core |
planned | ProtocolHandler trait, ALPN router, auth/identity, config |
alknet-ssh |
planned | SSH handler (russh), SOCKS5, port forwarding |
alknet-call |
planned | JSON-RPC call protocol (EventEnvelope framing) |
alknet-fs |
planned | Content-addressed file storage (iroh-blobs backend) |
alknet-sftp |
planned | SFTP handler (russh-sftp protocol core) |
alknet-git |
planned | Git smart protocol handler (gix) |
alknet-http |
planned | HTTP handler (axum, REST API, MCP) |
alknet-dns |
planned | DNS handler (hickory-proto, pkarr) |
alknet-msg |
planned | E2E encrypted messaging, mixnet support |
alknet |
planned | CLI binary (assembles and registers handlers) |
Documentation
- ALPN-as-service architecture — pivot proposal
- Cleanup plan — greenfield transition plan
- SDD process — spec-driven development process
- Research references — iroh, russh, russh-sftp deep dives
Reference implementation (previous architecture) is preserved at /workspace/@alkdev/alknet-main/.
License
Licensed under either of
- Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
- MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)
at your option.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in this work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.