| Filename | Latest commit message | Latest commit date |
|---|---|---|
|
All checks were successful
energy meters on real hardware / RAPL and NVML on jetson-hub (pull_request) Successful in 40s
standards workspaces build + test / jouleclaw-rs (pull_request) Successful in 3m12s
standards workspaces build + test / aac-rs (pull_request) Successful in 2m57s
standards workspaces build + test / joule-code-rs (pull_request) Successful in 1m23s
standards workspaces build + test / joulecontract-rs (pull_request) Successful in 55s
standards workspaces build + test / map-rs (pull_request) Successful in 56s
standards workspaces build + test / mesh-rs (pull_request) Successful in 53s
standards workspaces build + test / openpay (pull_request) Successful in 2m39s
standards workspaces build + test / jouledesk-rs (pull_request) Successful in 43s
standards workspaces build + test / macOS energy backends (pull_request) Successful in 4m4s
jouleclaw public surface must stay documented / Public-surface audit (jouleclaw-rs) (pull_request) Successful in 1m8s
standards workspaces build + test / arl-rs (pull_request) Successful in 1m18s
standards workspaces build + test / eoc-rs (pull_request) Successful in 2m3s
standards workspaces build + test / jcp-rs (pull_request) Successful in 1m38s
standards workspaces build + test / joule-ui-rs (pull_request) Successful in 1m12s
standards workspaces build + test / joulehook-rs (pull_request) Successful in 1m13s
standards workspaces build + test / proof-rs (pull_request) Successful in 1m19s
standards workspaces build + test / sense-rs (pull_request) Successful in 1m19s
standards workspaces build + test / sandbox-rs (pull_request) Successful in 2m25s
standards workspaces build + test / smart-byte-rs (pull_request) Successful in 4m3s
`the_production_probe_needs_every_package_to_keep_advancing` drove a fake two-socket tree from a writer thread on a 2 ms timer and expected the production probe to call it live inside one budget. On a loaded host the writer is starved, the samples stop changing, and the probe rightly judges the scripted tree frozen, so the test failed on load rather than on behaviour. The probe now has a `probe_stepping` form that runs a closure at the start of every tick, ahead of the burn. The real probe passes nothing. The test advances the tree there, so each interval the probe samples holds one advance however threads are scheduled, and no second thread exists. The timing claim is also restated as what a loaded host leaves certain: a live verdict comes no sooner than the rule's window and inside the probe's own cap, and a frozen verdict takes at least one budget. Verified on the Linux x86_64 host with `--features rapl,nvml`: the lib tests pass idle, and the RAPL tests pass on every one of 25 repeats under 40 busy loops. Co-Authored-By: Claude Opus 5.5 <[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)