memtrust-cli 0.3.0__tar.gz → 0.3.2__tar.gz

memtrust_cli-0.3.2/.gitignore

@@ -0,0 +1,34 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+.eggs/
+build/
+dist/
+.venv/
+venv/
+.env
+.env.*
+!.env.example
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.coverage
+.coverage.*
+htmlcov/
+coverage.xml
+*.log
+.DS_Store
+results/*/
+!results/.gitkeep
+node_modules/
+.idea/
+.vscode/
+.claude/
+uv.lock
+memtrust-report-*.json
+
+# Locally-downloaded uv binaries staged into the npm platform packages
+# before `npm publish` -- real binaries, ~50-70MB each across 6 platforms,
+# never meant to be committed or swept into the Python sdist. See
+# npm/platforms/*/README.md for how each one is populated at publish time.
+npm/platforms/*/bin/

memtrust_cli-0.3.0/README.md → memtrust_cli-0.3.2/PKG-INFO

@@ -1,3 +1,48 @@
+Metadata-Version: 2.4
+Name: memtrust-cli
+Version: 0.3.2
+Summary: Independent, reproducible benchmark harness for agent-memory backends (MemPalace, Mem0, Zep/Graphiti, OpenViking): runs LongMemEval, LoCoMo, contradiction-detection, and a dozen other evals against all four and publishes the raw logs, not vendor-curated numbers.
+Project-URL: Homepage, https://github.com/RudrenduPaul/memtrust
+Project-URL: Issues, https://github.com/RudrenduPaul/memtrust/issues
+Project-URL: Author - Rudrendu Paul, https://github.com/RudrenduPaul
+Project-URL: Author - Sourav Nandy, https://github.com/Sourav-nandy-ai
+Author: Rudrendu Paul, Sourav Nandy
+License: Apache-2.0
+License-File: LICENSE
+Keywords: agent-memory,ai-agents,benchmark,contradiction-detection,graphiti,llm-eval,llm-memory,locomo,longmemeval,mem0,memory-backend,mempalace,reproducible-benchmark,zep
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Software Development :: Testing
+Requires-Python: >=3.11
+Requires-Dist: click<9.0,>=8.1.7
+Requires-Dist: cryptography<49.0,>=48.0.1
+Requires-Dist: httpx<1.0,>=0.27.0
+Requires-Dist: rich<14.0,>=13.7.1
+Provides-Extra: dev
+Requires-Dist: mypy<2.0,>=1.10.0; extra == 'dev'
+Requires-Dist: pip-audit<3.0,>=2.7.3; extra == 'dev'
+Requires-Dist: pytest-cov<6.0,>=5.0.0; extra == 'dev'
+Requires-Dist: pytest-httpx<1.0,>=0.36.2; extra == 'dev'
+Requires-Dist: pytest<10.0,>=9.0.3; extra == 'dev'
+Requires-Dist: ruff>=0.5.0; extra == 'dev'
+Provides-Extra: mem0-direct
+Requires-Dist: boto3<2.0,>=1.34.0; extra == 'mem0-direct'
+Requires-Dist: elasticsearch<9.0,>=8.0.0; extra == 'mem0-direct'
+Requires-Dist: fastembed<1.0,>=0.3.1; extra == 'mem0-direct'
+Requires-Dist: google-genai<2.0,>=1.0.0; extra == 'mem0-direct'
+Requires-Dist: mem0ai<3.0,>=2.0.12; extra == 'mem0-direct'
+Requires-Dist: redis<6.0,>=5.0.0; extra == 'mem0-direct'
+Requires-Dist: redisvl<1.0,>=0.3.0; extra == 'mem0-direct'
+Requires-Dist: valkey<7.0,>=6.0.0; extra == 'mem0-direct'
+Provides-Extra: mempalace-direct
+Requires-Dist: mempalace<4.0,>=3.0.0; extra == 'mempalace-direct'
+Description-Content-Type: text/markdown
+
 # memtrust
 
 Standardized, reproducible benchmarks for agent-memory backends, run against the vendors, not
@@ -6,13 +51,15 @@ published by them.
 [![CI](https://github.com/RudrenduPaul/memtrust/actions/workflows/ci.yml/badge.svg)](https://github.com/RudrenduPaul/memtrust/actions/workflows/ci.yml)
 [![License: Apache-2.0](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](https://github.com/RudrenduPaul/memtrust/blob/main/LICENSE)
 [![Version](https://img.shields.io/badge/version-0.2.0-blue.svg)](https://github.com/RudrenduPaul/memtrust/blob/main/pyproject.toml)
-[![PyPI](https://img.shields.io/badge/pypi-memtrust-blue.svg)](https://pypi.org/project/memtrust/)
+[![PyPI](https://img.shields.io/badge/pypi-memtrust--cli-blue.svg)](https://pypi.org/project/memtrust-cli/)
 
 ```bash
-pip install memtrust
+pip install memtrust-cli
 memtrust run --backends mempalace,mem0,zep,openviking --eval all
 ```
 
+![Terminal recording of installing memtrust-cli with pip into a clean virtualenv, then running memtrust run against all four tracked backends with no credentials configured -- every backend reports SKIPPED and a JSON report is still written.](docs/demo.gif)
+
 (For contributing to this repo instead of just running it, see [Development](#development) --
 `pip install -e ".[dev]"` from a clone.)
 
@@ -24,7 +71,7 @@ memtrust run --backends mempalace,mem0,zep,openviking --eval all
 [The landscape](#the-landscape-verified-not-benchmarked) · [Benchmarks](#benchmarks) ·
 [GitHub Actions usage](#github-actions-usage) · [Self-host](#self-host) · [Install](#install) ·
 [Hosted layer](#what-a-hosted-trust-layer-would-add) · [Backend coverage](#backend-coverage) ·
-[Development](#development) · [License](#license) · [Success stories](#success-stories)
+[Development](#development) · [FAQ](#faq) · [License](#license) · [Success stories](#success-stories)
 
 ## Why this exists
 
@@ -177,6 +224,8 @@ a tested module and a MemPalace-specific capability, callable directly, not yet
 Every line above came straight from running `memtrust --help`, `memtrust run --help`, and
 `memtrust report --help` against this repo. Nothing here is invented.
 
+![Terminal recording of memtrust keygen generating an Ed25519 keypair, memtrust run --sign producing a signed receipt from a real run, and memtrust verify confirming the receipt's signature is valid.](docs/usage.gif)
+
 ## How this differs from trusting a vendor's own numbers
 
 Every backend memtrust tracks publishes its own benchmark numbers. None of them publish the same
@@ -322,7 +371,7 @@ jobs:
       - uses: actions/setup-python@v5
         with:
           python-version: "3.12"
-      - run: pip install memtrust
+      - run: pip install memtrust-cli
       - run: memtrust run --backends mempalace,mem0,zep,openviking --eval all --output leaderboard/data.json
         env:
           MEM0_API_KEY: ${{ secrets.MEM0_API_KEY }}
@@ -354,7 +403,7 @@ LongMemEval/LoCoMo datasets). Nothing leaves your machine unless you choose to p
 
 ### npx (agent-native)
 
-> **PyPI is live; the npm wrapper is publish-ready but not yet published.** `pip install memtrust`
+> **PyPI is live; the npm wrapper is publish-ready but not yet published.** `pip install memtrust-cli`
 > works today -- version 0.2.0 is on PyPI. The `memtrust-cli` npm package in `npm/memtrust-cli/` in
 > this repo is built, tested, and passes both `npm pack --dry-run` and `npm publish --dry-run`
 > cleanly (re-verified this session), but has not actually been pushed to the npm registry yet --
@@ -436,6 +485,76 @@ pip-audit
 `.pre-commit-config.yaml` wires ruff and mypy into `pre-commit` if you'd rather run these on every
 commit than remember to run them by hand.
 
+## FAQ
+
+**What is memtrust, and what actually makes it different from reading a vendor's own benchmark
+page?** It's a CLI harness that runs the same evals (LongMemEval, LoCoMo, and 14 others registered
+in `--eval`, including a contradiction-detection eval none of the four tracked backends publish a
+number for) against MemPalace, Mem0, Zep/Graphiti, and OpenViking, and prints the raw output rather
+than a curated summary. The differentiator isn't a proprietary scoring model; it's that nobody
+outside the vendor had previously run the same test, the same way, against every option, with the
+full methodology published alongside the code that produced it (`docs/methodology.md`). See "Why
+this exists" above for the MemPalace LongMemEval overclaim (mempalace/mempalace#27) that motivated
+the project.
+
+**Does memtrust support my platform, and what happens if it doesn't?** The npm wrapper
+(`memtrust-cli`) ships six platform-specific optional-dependency packages --
+`@memtrust-cli/darwin-x64`, `@memtrust-cli/darwin-arm64`, `@memtrust-cli/linux-x64`,
+`@memtrust-cli/linux-arm64`, `@memtrust-cli/win32-x64`, and `@memtrust-cli/win32-arm64` -- each
+bundling a verified `uv` binary for that exact platform (`npm/memtrust-cli/bin/memtrust.js`).
+`npm install` picks whichever one matches `process.platform`/`process.arch` at install time. On an
+unsupported combination (32-bit x86, or any platform outside that list), the wrapper exits with a
+clear `no prebuilt uv binary available for <platform>/<arch>` error instead of a silent failure. The
+underlying `memtrust` PyPI package itself only requires Python 3.11+, so `pip install memtrust-cli`
+remains the fallback path on any platform the npm wrapper doesn't cover.
+
+**Do I need a Python toolchain installed to use memtrust?** Not if you go through the npm wrapper.
+`npx memtrust-cli run ...` runs `uv tool run --from memtrust==<pinned version> memtrust <args>` under
+the hood, and `uv` provisions its own isolated Python interpreter and installs the exact pinned
+`memtrust` release from PyPI on first use, caching it after that. If you already have Python 3.11+,
+`pip install memtrust-cli` (the PyPI package name) works directly with no Node.js involved.
+
+**How does memtrust compare to a general-purpose LLM eval framework like RAGAS?** RAGAS evaluates
+RAG pipelines and other LLM applications with objective metrics and synthetic test-data generation;
+it has no memory-backend adapter abstraction and no eval built around a fact contradicting an
+earlier one, because that's not the problem it's built to solve. memtrust is narrower on purpose: it
+only tracks four named agent-memory backends and its non-recall evals (contradiction detection,
+compression/round-trip fidelity, temporal-KG boundary detection) exist specifically to test claims
+those four backends make about themselves. If you need broad RAG or prompt-evaluation coverage,
+RAGAS or a similar framework is the right tool; if you need to check whether a memory backend
+silently drops or overwrites a contradicted fact, memtrust is the one built for that question.
+
+**Why does `memtrust --version` print a version that doesn't match what pip says I installed?**
+This is a real, currently open limitation, not a hypothetical one -- installing `memtrust-cli` from
+PyPI into a clean virtualenv and running `pip show memtrust-cli` reports `0.3.0`, but
+`memtrust --version` prints `0.0.0+unknown`. The CLI's version string isn't reading the installed
+package metadata correctly. Until it's fixed, use `pip show memtrust-cli` (or
+`python3 -c "import importlib.metadata as m; print(m.version('memtrust-cli'))"`) to check what's
+actually installed, not the CLI's own `--version` output.
+
+**Has memtrust actually been run against a live memory backend, or is this all synthetic?** Both,
+and the README doesn't blur the line. The eval logic itself is proven against bundled synthetic
+fixtures and, for several adapters, real installed vendor packages with only the network boundary
+mocked (see the pytest coverage table above). But no eval in this repo has been run against a live
+MemPalace, Mem0, Zep, or OpenViking backend with real credentials as of this writing -- the
+"Benchmarks" section above states that directly rather than publishing a placeholder number. The
+"Backend coverage" table gives a per-adapter confidence level (high/medium/low) for exactly this
+reason; run it yourself against your own credentials to get a live-verified number.
+
+**Can I use memtrust commercially, and does it require attribution?** Yes. It's licensed under
+Apache License 2.0 (see `LICENSE`), which permits commercial use, modification, and distribution,
+and requires you to preserve the license and copyright notice and to note any changes you make to
+the code. It does not require you to open-source your own product just because you depend on
+memtrust.
+
+**How do I get real accuracy numbers instead of the "not yet measured" placeholder?** Set the
+credential environment variable for whichever backend you want to test
+(`MEMPALACE_STORAGE_PATH`, `MEM0_API_KEY`, `ZEP_API_KEY`, `OPENVIKING_API_KEY`, plus
+`MEMTRUST_JUDGE_API_KEY` for LLM-judged evals like LongMemEval and LoCoMo), then run
+`memtrust run --backends <name> --eval all` and `memtrust report <output-file>`. A backend with no
+credential configured prints `SKIPPED` and the run still completes and writes a valid JSON report --
+see "What it does" and "Benchmarks" above for the exact commands.
+
 ## License
 
 Apache 2.0. See `LICENSE`.
@@ -577,3 +696,4 @@ open, deliberately not counted as fixed: some point at real gaps this harness ge
 yet without a live vendor credential, and inflating a near-miss to PASS defeats the entire point of
 an independently-verified benchmark. See the confidence caveats throughout this README and in
 `docs/methodology.md` for exactly which claims rest on which kind of evidence.
+

memtrust_cli-0.3.0/PKG-INFO → memtrust_cli-0.3.2/README.md

@@ -1,48 +1,3 @@
-Metadata-Version: 2.4
-Name: memtrust-cli
-Version: 0.3.0
-Summary: Independent, reproducible benchmark harness for agent-memory backends
-Project-URL: Homepage, https://github.com/RudrenduPaul/memtrust
-Project-URL: Issues, https://github.com/RudrenduPaul/memtrust/issues
-Project-URL: Author - Rudrendu Paul, https://github.com/RudrenduPaul
-Project-URL: Author - Sourav Nandy, https://github.com/Sourav-nandy-ai
-Author: Rudrendu Paul, Sourav Nandy
-License: Apache-2.0
-License-File: LICENSE
-Keywords: agent,benchmark,eval,evaluation,llm,memory
-Classifier: Development Status :: 3 - Alpha
-Classifier: Intended Audience :: Developers
-Classifier: License :: OSI Approved :: Apache Software License
-Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.11
-Classifier: Programming Language :: Python :: 3.12
-Classifier: Programming Language :: Python :: 3.13
-Classifier: Topic :: Software Development :: Testing
-Requires-Python: >=3.11
-Requires-Dist: click<9.0,>=8.1.7
-Requires-Dist: cryptography<49.0,>=48.0.1
-Requires-Dist: httpx<1.0,>=0.27.0
-Requires-Dist: rich<14.0,>=13.7.1
-Provides-Extra: dev
-Requires-Dist: mypy<2.0,>=1.10.0; extra == 'dev'
-Requires-Dist: pip-audit<3.0,>=2.7.3; extra == 'dev'
-Requires-Dist: pytest-cov<6.0,>=5.0.0; extra == 'dev'
-Requires-Dist: pytest-httpx<1.0,>=0.36.2; extra == 'dev'
-Requires-Dist: pytest<10.0,>=9.0.3; extra == 'dev'
-Requires-Dist: ruff>=0.5.0; extra == 'dev'
-Provides-Extra: mem0-direct
-Requires-Dist: boto3<2.0,>=1.34.0; extra == 'mem0-direct'
-Requires-Dist: elasticsearch<9.0,>=8.0.0; extra == 'mem0-direct'
-Requires-Dist: fastembed<1.0,>=0.3.1; extra == 'mem0-direct'
-Requires-Dist: google-genai<2.0,>=1.0.0; extra == 'mem0-direct'
-Requires-Dist: mem0ai<3.0,>=2.0.12; extra == 'mem0-direct'
-Requires-Dist: redis<6.0,>=5.0.0; extra == 'mem0-direct'
-Requires-Dist: redisvl<1.0,>=0.3.0; extra == 'mem0-direct'
-Requires-Dist: valkey<7.0,>=6.0.0; extra == 'mem0-direct'
-Provides-Extra: mempalace-direct
-Requires-Dist: mempalace<4.0,>=3.0.0; extra == 'mempalace-direct'
-Description-Content-Type: text/markdown
-
 # memtrust
 
 Standardized, reproducible benchmarks for agent-memory backends, run against the vendors, not
@@ -51,13 +6,15 @@ published by them.
 [![CI](https://github.com/RudrenduPaul/memtrust/actions/workflows/ci.yml/badge.svg)](https://github.com/RudrenduPaul/memtrust/actions/workflows/ci.yml)
 [![License: Apache-2.0](https://img.shields.io/badge/license-Apache--2.0-blue.svg)](https://github.com/RudrenduPaul/memtrust/blob/main/LICENSE)
 [![Version](https://img.shields.io/badge/version-0.2.0-blue.svg)](https://github.com/RudrenduPaul/memtrust/blob/main/pyproject.toml)
-[![PyPI](https://img.shields.io/badge/pypi-memtrust-blue.svg)](https://pypi.org/project/memtrust/)
+[![PyPI](https://img.shields.io/badge/pypi-memtrust--cli-blue.svg)](https://pypi.org/project/memtrust-cli/)
 
 ```bash
-pip install memtrust
+pip install memtrust-cli
 memtrust run --backends mempalace,mem0,zep,openviking --eval all
 ```
 
+![Terminal recording of installing memtrust-cli with pip into a clean virtualenv, then running memtrust run against all four tracked backends with no credentials configured -- every backend reports SKIPPED and a JSON report is still written.](docs/demo.gif)
+
 (For contributing to this repo instead of just running it, see [Development](#development) --
 `pip install -e ".[dev]"` from a clone.)
 
@@ -69,7 +26,7 @@ memtrust run --backends mempalace,mem0,zep,openviking --eval all
 [The landscape](#the-landscape-verified-not-benchmarked) · [Benchmarks](#benchmarks) ·
 [GitHub Actions usage](#github-actions-usage) · [Self-host](#self-host) · [Install](#install) ·
 [Hosted layer](#what-a-hosted-trust-layer-would-add) · [Backend coverage](#backend-coverage) ·
-[Development](#development) · [License](#license) · [Success stories](#success-stories)
+[Development](#development) · [FAQ](#faq) · [License](#license) · [Success stories](#success-stories)
 
 ## Why this exists
 
@@ -222,6 +179,8 @@ a tested module and a MemPalace-specific capability, callable directly, not yet
 Every line above came straight from running `memtrust --help`, `memtrust run --help`, and
 `memtrust report --help` against this repo. Nothing here is invented.
 
+![Terminal recording of memtrust keygen generating an Ed25519 keypair, memtrust run --sign producing a signed receipt from a real run, and memtrust verify confirming the receipt's signature is valid.](docs/usage.gif)
+
 ## How this differs from trusting a vendor's own numbers
 
 Every backend memtrust tracks publishes its own benchmark numbers. None of them publish the same
@@ -367,7 +326,7 @@ jobs:
       - uses: actions/setup-python@v5
         with:
           python-version: "3.12"
-      - run: pip install memtrust
+      - run: pip install memtrust-cli
       - run: memtrust run --backends mempalace,mem0,zep,openviking --eval all --output leaderboard/data.json
         env:
           MEM0_API_KEY: ${{ secrets.MEM0_API_KEY }}
@@ -399,7 +358,7 @@ LongMemEval/LoCoMo datasets). Nothing leaves your machine unless you choose to p
 
 ### npx (agent-native)
 
-> **PyPI is live; the npm wrapper is publish-ready but not yet published.** `pip install memtrust`
+> **PyPI is live; the npm wrapper is publish-ready but not yet published.** `pip install memtrust-cli`
 > works today -- version 0.2.0 is on PyPI. The `memtrust-cli` npm package in `npm/memtrust-cli/` in
 > this repo is built, tested, and passes both `npm pack --dry-run` and `npm publish --dry-run`
 > cleanly (re-verified this session), but has not actually been pushed to the npm registry yet --
@@ -481,6 +440,76 @@ pip-audit
 `.pre-commit-config.yaml` wires ruff and mypy into `pre-commit` if you'd rather run these on every
 commit than remember to run them by hand.
 
+## FAQ
+
+**What is memtrust, and what actually makes it different from reading a vendor's own benchmark
+page?** It's a CLI harness that runs the same evals (LongMemEval, LoCoMo, and 14 others registered
+in `--eval`, including a contradiction-detection eval none of the four tracked backends publish a
+number for) against MemPalace, Mem0, Zep/Graphiti, and OpenViking, and prints the raw output rather
+than a curated summary. The differentiator isn't a proprietary scoring model; it's that nobody
+outside the vendor had previously run the same test, the same way, against every option, with the
+full methodology published alongside the code that produced it (`docs/methodology.md`). See "Why
+this exists" above for the MemPalace LongMemEval overclaim (mempalace/mempalace#27) that motivated
+the project.
+
+**Does memtrust support my platform, and what happens if it doesn't?** The npm wrapper
+(`memtrust-cli`) ships six platform-specific optional-dependency packages --
+`@memtrust-cli/darwin-x64`, `@memtrust-cli/darwin-arm64`, `@memtrust-cli/linux-x64`,
+`@memtrust-cli/linux-arm64`, `@memtrust-cli/win32-x64`, and `@memtrust-cli/win32-arm64` -- each
+bundling a verified `uv` binary for that exact platform (`npm/memtrust-cli/bin/memtrust.js`).
+`npm install` picks whichever one matches `process.platform`/`process.arch` at install time. On an
+unsupported combination (32-bit x86, or any platform outside that list), the wrapper exits with a
+clear `no prebuilt uv binary available for <platform>/<arch>` error instead of a silent failure. The
+underlying `memtrust` PyPI package itself only requires Python 3.11+, so `pip install memtrust-cli`
+remains the fallback path on any platform the npm wrapper doesn't cover.
+
+**Do I need a Python toolchain installed to use memtrust?** Not if you go through the npm wrapper.
+`npx memtrust-cli run ...` runs `uv tool run --from memtrust==<pinned version> memtrust <args>` under
+the hood, and `uv` provisions its own isolated Python interpreter and installs the exact pinned
+`memtrust` release from PyPI on first use, caching it after that. If you already have Python 3.11+,
+`pip install memtrust-cli` (the PyPI package name) works directly with no Node.js involved.
+
+**How does memtrust compare to a general-purpose LLM eval framework like RAGAS?** RAGAS evaluates
+RAG pipelines and other LLM applications with objective metrics and synthetic test-data generation;
+it has no memory-backend adapter abstraction and no eval built around a fact contradicting an
+earlier one, because that's not the problem it's built to solve. memtrust is narrower on purpose: it
+only tracks four named agent-memory backends and its non-recall evals (contradiction detection,
+compression/round-trip fidelity, temporal-KG boundary detection) exist specifically to test claims
+those four backends make about themselves. If you need broad RAG or prompt-evaluation coverage,
+RAGAS or a similar framework is the right tool; if you need to check whether a memory backend
+silently drops or overwrites a contradicted fact, memtrust is the one built for that question.
+
+**Why does `memtrust --version` print a version that doesn't match what pip says I installed?**
+This is a real, currently open limitation, not a hypothetical one -- installing `memtrust-cli` from
+PyPI into a clean virtualenv and running `pip show memtrust-cli` reports `0.3.0`, but
+`memtrust --version` prints `0.0.0+unknown`. The CLI's version string isn't reading the installed
+package metadata correctly. Until it's fixed, use `pip show memtrust-cli` (or
+`python3 -c "import importlib.metadata as m; print(m.version('memtrust-cli'))"`) to check what's
+actually installed, not the CLI's own `--version` output.
+
+**Has memtrust actually been run against a live memory backend, or is this all synthetic?** Both,
+and the README doesn't blur the line. The eval logic itself is proven against bundled synthetic
+fixtures and, for several adapters, real installed vendor packages with only the network boundary
+mocked (see the pytest coverage table above). But no eval in this repo has been run against a live
+MemPalace, Mem0, Zep, or OpenViking backend with real credentials as of this writing -- the
+"Benchmarks" section above states that directly rather than publishing a placeholder number. The
+"Backend coverage" table gives a per-adapter confidence level (high/medium/low) for exactly this
+reason; run it yourself against your own credentials to get a live-verified number.
+
+**Can I use memtrust commercially, and does it require attribution?** Yes. It's licensed under
+Apache License 2.0 (see `LICENSE`), which permits commercial use, modification, and distribution,
+and requires you to preserve the license and copyright notice and to note any changes you make to
+the code. It does not require you to open-source your own product just because you depend on
+memtrust.
+
+**How do I get real accuracy numbers instead of the "not yet measured" placeholder?** Set the
+credential environment variable for whichever backend you want to test
+(`MEMPALACE_STORAGE_PATH`, `MEM0_API_KEY`, `ZEP_API_KEY`, `OPENVIKING_API_KEY`, plus
+`MEMTRUST_JUDGE_API_KEY` for LLM-judged evals like LongMemEval and LoCoMo), then run
+`memtrust run --backends <name> --eval all` and `memtrust report <output-file>`. A backend with no
+credential configured prints `SKIPPED` and the run still completes and writes a valid JSON report --
+see "What it does" and "Benchmarks" above for the exact commands.
+
 ## License
 
 Apache 2.0. See `LICENSE`.
@@ -622,3 +651,4 @@ open, deliberately not counted as fixed: some point at real gaps this harness ge
 yet without a live vendor credential, and inflating a near-miss to PASS defeats the entire point of
 an independently-verified benchmark. See the confidence caveats throughout this README and in
 `docs/methodology.md` for exactly which claims rest on which kind of evidence.
+

{memtrust_cli-0.3.0 → memtrust_cli-0.3.2}/npm/memtrust-cli/package.json

@@ -1,6 +1,6 @@
 {
   "name": "memtrust-cli",
-  "version": "0.3.0",
+  "version": "0.3.2",
   "description": "npx-runnable wrapper for the memtrust agent-memory benchmark harness. Bootstraps a verified copy of Astral's uv to fetch and run memtrust from PyPI -- no Python toolchain to provision by hand.",
   "bin": { "memtrust-cli": "bin/memtrust.js", "memtrust": "bin/memtrust.js" },
   "files": ["bin/", "LICENSE", "README.md"],

memtrust_cli-0.3.2/npm/platforms/@memtrust-cli-darwin-arm64/README.md

@@ -0,0 +1,25 @@
+# @memtrust-cli/darwin-arm64
+
+macOS arm64 binary for [memtrust-cli](https://github.com/RudrenduPaul/memtrust)'s npx wrapper.
+
+This package is not meant to be installed directly. Install the main
+[`memtrust-cli`](https://www.npmjs.com/package/memtrust-cli) package instead; it
+pulls in this platform binary automatically through `optionalDependencies` based
+on your OS and CPU architecture.
+
+## What's inside
+
+A genuine, SHA-256-verified copy of [Astral's uv](https://github.com/astral-sh/uv),
+fetched from uv's own GitHub release 0.11.28 at publish time. uv is dual-licensed
+MIT OR Apache-2.0; both license texts are reproduced in this package's LICENSE
+file. Not affiliated with or endorsed by Astral Software Inc.
+
+## Documentation
+
+Full documentation, usage, and benchmarks live in the main
+[memtrust repository](https://github.com/RudrenduPaul/memtrust).
+
+## License
+
+Apache-2.0 for memtrust's own wrapper code. See LICENSE for the bundled uv
+binary's MIT OR Apache-2.0 terms.

{memtrust_cli-0.3.0 → memtrust_cli-0.3.2}/npm/platforms/@memtrust-cli-darwin-arm64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@memtrust-cli/darwin-arm64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "macOS arm64 binary for memtrust's npx wrapper. Bundles a genuine, SHA-256-verified copy of Astral's uv (github.com/astral-sh/uv), dual-licensed MIT OR Apache-2.0, fetched from uv's own GitHub release 0.11.28 at publish time. Not affiliated with or endorsed by Astral Software Inc.",
   "os": ["darwin"],
   "cpu": ["arm64"],

memtrust_cli-0.3.2/npm/platforms/@memtrust-cli-darwin-x64/README.md

@@ -0,0 +1,25 @@
+# @memtrust-cli/darwin-x64
+
+macOS x64 binary for [memtrust-cli](https://github.com/RudrenduPaul/memtrust)'s npx wrapper.
+
+This package is not meant to be installed directly. Install the main
+[`memtrust-cli`](https://www.npmjs.com/package/memtrust-cli) package instead; it
+pulls in this platform binary automatically through `optionalDependencies` based
+on your OS and CPU architecture.
+
+## What's inside
+
+A genuine, SHA-256-verified copy of [Astral's uv](https://github.com/astral-sh/uv),
+fetched from uv's own GitHub release 0.11.28 at publish time. uv is dual-licensed
+MIT OR Apache-2.0; both license texts are reproduced in this package's LICENSE
+file. Not affiliated with or endorsed by Astral Software Inc.
+
+## Documentation
+
+Full documentation, usage, and benchmarks live in the main
+[memtrust repository](https://github.com/RudrenduPaul/memtrust).
+
+## License
+
+Apache-2.0 for memtrust's own wrapper code. See LICENSE for the bundled uv
+binary's MIT OR Apache-2.0 terms.

{memtrust_cli-0.3.0 → memtrust_cli-0.3.2}/npm/platforms/@memtrust-cli-darwin-x64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@memtrust-cli/darwin-x64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "macOS x64 binary for memtrust's npx wrapper. Bundles a genuine, SHA-256-verified copy of Astral's uv (github.com/astral-sh/uv), dual-licensed MIT OR Apache-2.0, fetched from uv's own GitHub release 0.11.28 at publish time. Not affiliated with or endorsed by Astral Software Inc.",
   "os": ["darwin"],
   "cpu": ["x64"],

memtrust_cli-0.3.2/npm/platforms/@memtrust-cli-linux-arm64/README.md

@@ -0,0 +1,25 @@
+# @memtrust-cli/linux-arm64
+
+Linux arm64 binary for [memtrust-cli](https://github.com/RudrenduPaul/memtrust)'s npx wrapper.
+
+This package is not meant to be installed directly. Install the main
+[`memtrust-cli`](https://www.npmjs.com/package/memtrust-cli) package instead; it
+pulls in this platform binary automatically through `optionalDependencies` based
+on your OS and CPU architecture.
+
+## What's inside
+
+A genuine, SHA-256-verified copy of [Astral's uv](https://github.com/astral-sh/uv),
+fetched from uv's own GitHub release 0.11.28 at publish time. uv is dual-licensed
+MIT OR Apache-2.0; both license texts are reproduced in this package's LICENSE
+file. Not affiliated with or endorsed by Astral Software Inc.
+
+## Documentation
+
+Full documentation, usage, and benchmarks live in the main
+[memtrust repository](https://github.com/RudrenduPaul/memtrust).
+
+## License
+
+Apache-2.0 for memtrust's own wrapper code. See LICENSE for the bundled uv
+binary's MIT OR Apache-2.0 terms.

{memtrust_cli-0.3.0 → memtrust_cli-0.3.2}/npm/platforms/@memtrust-cli-linux-arm64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@memtrust-cli/linux-arm64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "Linux arm64 binary for memtrust's npx wrapper. Bundles a genuine, SHA-256-verified copy of Astral's uv (github.com/astral-sh/uv), dual-licensed MIT OR Apache-2.0, fetched from uv's own GitHub release 0.11.28 at publish time. Not affiliated with or endorsed by Astral Software Inc.",
   "os": ["linux"],
   "cpu": ["arm64"],

memtrust_cli-0.3.2/npm/platforms/@memtrust-cli-linux-x64/README.md

@@ -0,0 +1,25 @@
+# @memtrust-cli/linux-x64
+
+Linux x64 binary for [memtrust-cli](https://github.com/RudrenduPaul/memtrust)'s npx wrapper.
+
+This package is not meant to be installed directly. Install the main
+[`memtrust-cli`](https://www.npmjs.com/package/memtrust-cli) package instead; it
+pulls in this platform binary automatically through `optionalDependencies` based
+on your OS and CPU architecture.
+
+## What's inside
+
+A genuine, SHA-256-verified copy of [Astral's uv](https://github.com/astral-sh/uv),
+fetched from uv's own GitHub release 0.11.28 at publish time. uv is dual-licensed
+MIT OR Apache-2.0; both license texts are reproduced in this package's LICENSE
+file. Not affiliated with or endorsed by Astral Software Inc.
+
+## Documentation
+
+Full documentation, usage, and benchmarks live in the main
+[memtrust repository](https://github.com/RudrenduPaul/memtrust).
+
+## License
+
+Apache-2.0 for memtrust's own wrapper code. See LICENSE for the bundled uv
+binary's MIT OR Apache-2.0 terms.

{memtrust_cli-0.3.0 → memtrust_cli-0.3.2}/npm/platforms/@memtrust-cli-linux-x64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@memtrust-cli/linux-x64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "Linux x64 binary for memtrust's npx wrapper. Bundles a genuine, SHA-256-verified copy of Astral's uv (github.com/astral-sh/uv), dual-licensed MIT OR Apache-2.0, fetched from uv's own GitHub release 0.11.28 at publish time. Not affiliated with or endorsed by Astral Software Inc.",
   "os": ["linux"],
   "cpu": ["x64"],

memtrust_cli-0.3.2/npm/platforms/@memtrust-cli-win32-arm64/README.md

@@ -0,0 +1,25 @@
+# @memtrust-cli/win32-arm64
+
+Windows arm64 binary for [memtrust-cli](https://github.com/RudrenduPaul/memtrust)'s npx wrapper.
+
+This package is not meant to be installed directly. Install the main
+[`memtrust-cli`](https://www.npmjs.com/package/memtrust-cli) package instead; it
+pulls in this platform binary automatically through `optionalDependencies` based
+on your OS and CPU architecture.
+
+## What's inside
+
+A genuine, SHA-256-verified copy of [Astral's uv](https://github.com/astral-sh/uv),
+fetched from uv's own GitHub release 0.11.28 at publish time. uv is dual-licensed
+MIT OR Apache-2.0; both license texts are reproduced in this package's LICENSE
+file. Not affiliated with or endorsed by Astral Software Inc.
+
+## Documentation
+
+Full documentation, usage, and benchmarks live in the main
+[memtrust repository](https://github.com/RudrenduPaul/memtrust).
+
+## License
+
+Apache-2.0 for memtrust's own wrapper code. See LICENSE for the bundled uv
+binary's MIT OR Apache-2.0 terms.

{memtrust_cli-0.3.0 → memtrust_cli-0.3.2}/npm/platforms/@memtrust-cli-win32-arm64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@memtrust-cli/win32-arm64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "Windows arm64 binary for memtrust's npx wrapper. Bundles a genuine, SHA-256-verified copy of Astral's uv (github.com/astral-sh/uv), dual-licensed MIT OR Apache-2.0, fetched from uv's own GitHub release 0.11.28 at publish time. Not affiliated with or endorsed by Astral Software Inc.",
   "os": ["win32"],
   "cpu": ["arm64"],

memtrust_cli-0.3.2/npm/platforms/@memtrust-cli-win32-x64/README.md

@@ -0,0 +1,25 @@
+# @memtrust-cli/win32-x64
+
+Windows x64 binary for [memtrust-cli](https://github.com/RudrenduPaul/memtrust)'s npx wrapper.
+
+This package is not meant to be installed directly. Install the main
+[`memtrust-cli`](https://www.npmjs.com/package/memtrust-cli) package instead; it
+pulls in this platform binary automatically through `optionalDependencies` based
+on your OS and CPU architecture.
+
+## What's inside
+
+A genuine, SHA-256-verified copy of [Astral's uv](https://github.com/astral-sh/uv),
+fetched from uv's own GitHub release 0.11.28 at publish time. uv is dual-licensed
+MIT OR Apache-2.0; both license texts are reproduced in this package's LICENSE
+file. Not affiliated with or endorsed by Astral Software Inc.
+
+## Documentation
+
+Full documentation, usage, and benchmarks live in the main
+[memtrust repository](https://github.com/RudrenduPaul/memtrust).
+
+## License
+
+Apache-2.0 for memtrust's own wrapper code. See LICENSE for the bundled uv
+binary's MIT OR Apache-2.0 terms.

{memtrust_cli-0.3.0 → memtrust_cli-0.3.2}/npm/platforms/@memtrust-cli-win32-x64/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@memtrust-cli/win32-x64",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "Windows x64 binary for memtrust's npx wrapper. Bundles a genuine, SHA-256-verified copy of Astral's uv (github.com/astral-sh/uv), dual-licensed MIT OR Apache-2.0, fetched from uv's own GitHub release 0.11.28 at publish time. Not affiliated with or endorsed by Astral Software Inc.",
   "os": ["win32"],
   "cpu": ["x64"],