| Filename | Latest commit message | Latest commit date |
|---|---|---|
|
All checks were successful
jouleclaw public surface must stay documented / Public-surface audit (jouleclaw-rs) (pull_request) Successful in 1m2s
standards workspaces build + test / which workspaces this change touches (pull_request) Successful in 6s
repository checks / lockfiles, tests in CI, workflow hygiene, spec and section references, public text (pull_request) Successful in 46s
standards workspaces build + test / arl-rs (arl/arl-rs, arl-rs, linux) (pull_request) Successful in 1m16s
standards workspaces build + test / jcp-rs (jcp-conformance, jcp/jcp-rs, -p jcp-energy --features nvml, jcp-rs, linux) (pull_request) Successful in 1m53s
standards workspaces build + test / eoc-rs (eoc/eoc-rs, -p eoc-meter --features cuda, eoc-rs, linux) (pull_request) Successful in 2m9s
standards workspaces build + test / joule-ui-rs (joule-ui/joule-ui-rs, joule-ui-rs, linux) (pull_request) Successful in 1m14s
standards workspaces build + test / joulehook-rs (joulehook/joulehook-rs, joulehook-rs, linux) (pull_request) Successful in 1m11s
standards workspaces build + test / pool-rs (pool/pool-rs, pool-rs, linux) (pull_request) Successful in 1m8s
standards workspaces build + test / proof-rs (proof-conformance, proof/proof-rs, proof-rs, linux, true) (pull_request) Successful in 1m28s
standards workspaces build + test / sense-rs (sense/sense-rs, sense-rs, linux) (pull_request) Successful in 1m16s
standards workspaces build + test / sandbox-rs (sandbox-conformance, sandbox/sandbox-rs, sandbox-rs, linux) (pull_request) Successful in 2m27s
standards workspaces build + test / smart-byte-rs (smart-byte/smart-byte-rs, smart-byte-rs, linux) (pull_request) Successful in 4m10s
standards workspaces build + test / jouleclaw-rs (-nodebug, jouleclaw/jouleclaw-rs, -p jouleclaw-energy --features rapl,nvml, jouleclaw-rs, linux-arm64, -C target-feature=+fp16) (pull_request) Successful in 6m50s
standards workspaces build + test / joulecontract-rs (joulecontract/joulecontract-rs, joulecontract-rs, linux-arm64, true) (pull_request) Successful in 1m30s
standards workspaces build + test / map-rs (map-conformance, map/map-rs, map-rs, linux-arm64) (pull_request) Successful in 1m28s
standards workspaces build + test / aac-rs (aac-conformance, aac/aac-rs, aac-rs, linux-arm64) (pull_request) Successful in 5m17s
standards workspaces build + test / joule-code-rs (joule-code/joule-code-rs, joule-code-rs, linux-arm64, true) (pull_request) Successful in 5m10s
standards workspaces build + test / mesh-rs (mesh/mesh-rs, -p mesh-engine-gguf -p mesh-openai --features mesh-engine-gguf/gguf,mesh-openai/server, mesh-rs, linux-arm64) (pull_request) Successful in 2m24s
standards workspaces build + test / openpay (openpay, openpay, linux-arm64) (pull_request) Successful in 6m16s
standards workspaces build + test / macOS energy backends (pull_request) Successful in 8m16s
standards workspaces build + test / jouledesk-rs (jouledesk/jouledesk-rs, jouledesk-rs, macos) (pull_request) Successful in 2m52s
12 -> 0, in both the default and `rayon fabric-meter` configurations, and the crate joins the `-D warnings` clippy step in standards-build.yml. The last dozen needed judgement rather than a script: - two `from_str` methods that are not FromStr-shaped (one returns Option, the other is infallible) renamed to `parse`. Both are file-local, so there is no caller churn -- except that my check for callers missed mlx.rs, which reaches SafetensorsDtype through a type ALIAS, so `SafetensorsDtype::from_str(` never matched. Renaming wants a grep for the method name across the workspace, not for the type names you happen to know. - `assert!(FLASH_KV_CHUNK.is_power_of_two())` and `>= 64` moved out of a `#[test]` into `const _: () = assert!(..)` beside the constant. These are properties of a constant; a test only holds them if someone runs it. - the `kv_quant_tests` module moved to the end of kv_cache_inplace.rs, because 300 lines of real items sat after it. - an example discarded the `Result` from `matmul_bf16_threaded`, so a failure would have been timed as a success. - `needless_range_loop` allowed per-file in three benches and one test: those are separate crates, so src/lib.rs's decision does not reach them. - checked_div for two guarded divisions, and one unused `mut`. Two near-misses worth recording: - `fabric_dispatch.rs` calls `ledger_record` twice with identical arguments, which reads exactly like a copy-paste bug. The comment three lines above says attention is "recorded as two GEMM-equivalents" so the measured µJ/GFLOP table prices it. Deliberate. Left alone. - clippy's `chunks_exact_mut` suggestion is `as_chunks_mut::<N>().0.iter()` -- a non-mutable iterator over a mutable chunk. It does not compile, and only the fabric-meter configuration contained one, so the default build stayed green while that example was broken. Checking both configurations is what caught it. The gate covers default features only. Under `rayon fabric-meter` this crate is also clean, but clippy reports jouleclaw-backend-apple through the path dependency, and that crate has 43 warnings -- 27 `unused_unsafe` and 15 `unused_imports` that are unused in THAT configuration and needed in others. Those want cfg work, not deletion, which is the trap #373 already paid for once. Verified: four lint specs pass, 343 loader-gguf tests, 77 omni tests. Co-Authored-By: Claude Opus 5 <[email protected]> |
||
| .. | ||
| conformance/v1 | ||
| jouleclaw-rs | ||
| scales | ||
| vocab | ||
| LICENSE | ||
| LICENSE-SPEC | ||
| README.md | ||
| ROADMAP.md | ||
| SPEC.md | ||
JouleClaw — Energy-Optimized AI Runtime
The substrate for AI that costs Joules, not tokens.
Capability per Joule, not capability per parameter.
What JouleClaw is
JouleClaw is a pure-Rust, omni-modal AI runtime + harness + agentic layer governed by Transaction Science as an open standard (5th in the open-standards quartet, alongside OpenPay, Smart Byte, EOC, and WAI).
Every operation JouleClaw performs walks an explicit energy-tiered cascade, top to bottom, stopping at the first stage that closes the query:
| Tier | Name | Cost class | What runs here |
|---|---|---|---|
| L0 | Cache | picojoules | Content-addressed cache hits |
| L1 | Lawful | nanojoules | Deterministic, pre-compiled primitive (text/code only) |
| L2 | Embed | sub-millijoules | Matryoshka embeddings + nearest-neighbor + hybrid search |
| L3 | Model | joules | Local SSM / ternary / multimodal / diffusion / etc. |
| L4 | Wire | tens of joules | Remote frontier RPC (escape hatch only) |
L1 (Lawful) is only meaningful for text and code modalities. For image, audio, video and 3D generation, the cascade collapses to L0 / L2 / L3 / L4 — the energy savings come from caching plus dispatch to the cheapest model that meets the quality bar.
Why JouleClaw exists
The mainstream "agentic AI" ecosystem treats inference as the entry point. Every state shift, every tool call, every routing decision runs through a 70-billion-parameter stochastic engine, burning kilojoules to do work that deterministic compute or a hot cache could resolve for microjoules.
JouleClaw inverts the default. Inference is the last resort, not the first. Frozen model weights are less trustworthy than current world state; fresh, provenance-stamped retrieval beats synthesis from training cuts. Every operation is metered in microjoules where the hardware permits it, and in millijoules where it doesn't — the spec is honest about the floor on every platform.
The thermodynamic circuit breaker is the safety mechanism. Set an energy budget; when measured consumption exceeds the budget, the breaker trips and the operation halts. Hallucination loops cannot survive an honest joule ledger.
What's here
| Path | Contents |
|---|---|
SPEC.md |
JouleClaw v1 spec — cascade tiers, energy contract, receipt |
CHARTER.md |
Stewardship pattern (shared across open-standards) |
jouleclaw-rs/ |
Apache-2.0 Rust reference implementation |
spec/ |
Tier specifications, energy provenance, receipt format |
conformance/v1/ |
Signed wire vectors for cross-implementation verification |
Reference implementation
jouleclaw-rs/ is the Apache-2.0 reference implementation in Rust. It
runs on Apple Silicon (M3/M4/M5 with UMA + Metal 4 Tensor APIs), AMD Strix
Halo / Ryzen AI Max, Intel Lunar Lake, NVIDIA Jetson, and discrete
NVIDIA / AMD GPUs.
It is omni-modal: text, code, vision, audio, video, image diffusion, 3D Gaussian splatting, and cross-modal fusion all dispatch through the same energy-tiered cascade.
The MCP tool surface is a JCP consumer: jouleclaw-jcp wraps any
jouleclaw-mcp metered tool so every dispatch is decided by the JCP runtime
against a signed, metered Grant before a byte reaches the tool — the tool's
identity is the content-addressed resource, the call's energy estimate is
debited from the grant's joule budget, and every decision (allow and deny
alike) seals as a JCR-1 receipt. An injected instruction can name an ungranted
tool; it cannot make the gate dispatch it.
Status
v0.1.0 reference implementation in active development. The spec is draft status. Cross-implementation conformance vectors will land with v1.0.0.
License
- Code: Apache-2.0 (see
LICENSE) - Spec text: CC-BY-4.0 (see
LICENSE-SPEC)