| Filename | Latest commit message | Latest commit date |
|---|---|---|
35e4ea0cb7
All checks were successful
energy meters on real hardware / RAPL and NVML on jetson-hub (pull_request) Successful in 33s
jouleclaw public surface must stay documented / Public-surface audit (jouleclaw-rs) (pull_request) Successful in 58s
standards workspaces build + test / arl-rs (pull_request) Successful in 1m7s
standards workspaces build + test / eoc-rs (pull_request) Successful in 1m53s
standards workspaces build + test / jcp-rs (pull_request) Successful in 1m27s
standards workspaces build + test / joule-ui-rs (pull_request) Successful in 1m8s
standards workspaces build + test / joulehook-rs (pull_request) Successful in 1m9s
standards workspaces build + test / jouleclaw-rs (pull_request) Successful in 1m52s
standards workspaces build + test / aac-rs (pull_request) Successful in 1m2s
standards workspaces build + test / proof-rs (pull_request) Successful in 1m14s
standards workspaces build + test / joule-code-rs (pull_request) Successful in 50s
standards workspaces build + test / joulecontract-rs (pull_request) Successful in 33s
standards workspaces build + test / map-rs (pull_request) Successful in 32s
standards workspaces build + test / sandbox-rs (pull_request) Successful in 2m10s
standards workspaces build + test / mesh-rs (pull_request) Successful in 28s
standards workspaces build + test / sense-rs (pull_request) Successful in 1m5s
standards workspaces build + test / openpay (pull_request) Successful in 1m34s
standards workspaces build + test / smart-byte-rs (pull_request) Successful in 3m37s
standards workspaces build + test / jouledesk-rs (pull_request) Successful in 3m55s
standards workspaces build + test / macOS energy backends (pull_request) Successful in 9m18s
Two tests failed on the dev Mac, which is also the macOS CI runner, at
load averages of 101-179. Neither failure was in the meters:
- eoc-meter `the_probe_burns_its_tick_of_thread_cpu_between_samples`.
A tick burned only 9.6 ms of the thread's CPU, under the 12.5 ms
floor, because its 500 ms wall cap cut it short. The test's own
message names that cap as the one thing allowed to shorten a tick,
but the test could not tell it apart: ProbeRun recorded no tick's
wall time. It now records each tick's wall time (`tick_walls`;
recorded, never decided on). A tick that ran its whole wall cap is
accepted as cut short, but only if it still spun, burning at least
1 ms of CPU. Any other tick must burn at least half its 25 ms, as
before.
- jouleclaw-energy `process_cpu`. The `/bin/sh` child, run for a fixed
300 ms, got 0.076 J of CPU against a 0.2 J sanity floor (load average
170). It failed 3 of 3 runs. The child now runs until its own counter
reads 0.2 J (at least 300 ms, at most 5 s), as eoc-meter's equivalent
test already does. The parent's window still spans the child's whole
run.
Both pass at a load average of 333. Both tick mutations are caught: a
tick that sleeps through its wall cap ("ran its whole 504.6ms wall cap
and burned only 87.375µs ... it waited, not spun"), and one that sleeps
25 ms.
Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
||
| .. | ||
| conformance/v1 | ||
| jouleclaw-rs | ||
| scales | ||
| vocab | ||
| LICENSE | ||
| 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
spec/LICENSE)