Implicit Core 1.0.0rc2 — initial public preview

Implicit Core 1.0.0rc2 is the initial public preview, licensed under Apache-2.0.

Implicit is the experience layer for AI agents. Virtualize large agent environments and materialize only the state each experience actually needs. Keep your existing agent, framework and evaluator.

RC2 adds bounded local stdio MCP, eight tools, adapter examples, reusable skill/plugin source and crawlable documentation. Runtime storage/execution semantics are inherited from rc1. The public preview includes a comment-only correction for cross-platform type checking.

The preserved rc1 benchmark recorded 155/155 equivalent cases, 93.88% aggregate, 94.40% mean, and 99.55% median retained serialized/materialized-state reduction, with approximately +0.554 seconds mean full-pipeline latency per case. These bytes are not RAM. Native replay assets are excluded; the public toy and sanitized numeric summary have separate scope. No allocator, learning, speed, SOTA or universal superiority claim is made.

Python 3.11–3.14 on Linux and Windows passed the public CI matrix. The canonical wheel is byte-reproducible from the sdist with SOURCE_DATE_EPOCH=1600000000.

Install the verified GitHub artifact in a fresh virtual environment:

python -m pip install https://github.com/zeitcow/implicit-core/releases/download/v1.0.0rc2/implicit_ai-1.0.0rc2-py3-none-any.whl
implicit --version
implicit --help
implicit demo
implicit benchmark

Avoid installing the unrelated implicit distribution in the same environment because it shares the import namespace.

Documentation · Benchmarks · Adapters · MCP

PyPI publication is pending the owner's authenticated trusted-publisher setup. Once published, the intended index command is python -m pip install --pre implicit-ai. The GitHub wheel is available independently.

Plugin source and a repository marketplace are public. Universal directory acceptance, desktop installation and hosted ChatGPT connectivity remain unverified. Search discovery assets are ready; actual indexing remains unknown.

Canonical package SHA-256:

• implicit_ai-1.0.0rc2-py3-none-any.whl: 7969e3ed4d9bfb393c2205574ec427b0c4d428b53490dd88ddbacca318e469ef
• implicit_ai-1.0.0rc2.tar.gz: 9253f2dcb29d7377c15ee9f821660ff67a10addba45675c0d126a443de9c6e3c

All attached assets are covered by checksums.txt. This is a release candidate, not stable 1.0.0.