Merge pull request #5 from hermes-labs-ai/codex/quickthink-v0.2.1-finalize

release: v0.2.1

Author: roli-lpci

.zenodo.json

@@ -1,6 +1,6 @@
 {
   "upload_type": "software",
-  "version": "0.2.0",
+  "version": "0.2.1",
   "title": "quickthink: local-first inference control layer for small LLMs",
   "description": "quickthink is a local-first inference control layer that aims to give small, on-device language models more reliable structured output by inserting a compressed plan-then-answer scaffold with latency-aware routing. Its routing, plan-grammar, and validation logic run as plain-Python middleware (no extra model of its own) in front of Ollama-backed calls, offering three execution modes (lite, two_pass, direct), preset routing profiles, a strict plan grammar, and a canonical run/judge/validate/report eval harness for reproducible local quality checks. Built by Hermes Labs as part of its open-source reliability layer for AI systems.",
   "creators": [

CHANGELOG.md

@@ -6,9 +6,15 @@ Format follows Keep a Changelog principles and semantic versioning.
 
 ## [Unreleased]
 
+## [0.2.1] - 2026-08-04
+
+### Added
+- `quickthink --version` reports the installed package version.
+
 ### Changed
-- Document the copyable, tagged-source install path to use before the first PyPI release.
-- Prepare tag-bound PyPI trusted publishing; no package has been uploaded.
+- Prepare the first PyPI publication through a tag-bound GitHub Actions trusted-publisher workflow. The workflow verifies that the tag and package version match, builds distributions, and checks them before the protected PyPI environment can publish.
+- Make source installation before publication and package installation after publication explicit in the quickstart.
+- Synchronize package, citation, and Zenodo release metadata for `0.2.1`.
 
 ## [0.2.0] - 2026-02-27
 ### Added

CITATION.cff

@@ -15,8 +15,8 @@ authors:
     given-names: "Rolando"
     orcid: "https://orcid.org/0009-0005-4896-1112"
     affiliation: "Hermes Labs"
-version: "0.2.0"
-date-released: "2026-02-27"
+version: "0.2.1"
+date-released: "2026-08-04"
 license: Apache-2.0
 repository-code: "https://github.com/hermes-labs-ai/quickthink"
 url: "https://github.com/hermes-labs-ai/quickthink"

README.md

@@ -61,8 +61,8 @@ Small/local models are fast but often underperform on multi-step tasks.
 Prerequisite: install and start [Ollama](https://ollama.com/) locally.
 
 ```bash
-# 1) Install the current tagged source (quickthink is not on PyPI yet)
-python -m pip install "quickthink @ git+https://github.com/hermes-labs-ai/quickthink.git@v0.2.0"
+# 1) Before the first PyPI publication, install the current source
+python -m pip install "quickthink @ git+https://github.com/hermes-labs-ai/quickthink.git"
 
 # 2) Pull one supported model
 ollama pull qwen2.5:1.5b
@@ -73,6 +73,12 @@ quickthink ask "Give me a 3-step plan to learn SQL basics" --model qwen2.5:1.5b
 
 If this command works, your local setup is ready.
 
+After `v0.2.1` has been published to PyPI, install the release instead:
+
+```bash
+python -m pip install "quickthink==0.2.1"
+```
+
 For development, clone the repository and install the editable development extras:
 
 ```bash

docs/release/RELEASE_NOTES_DRAFT.md

@@ -1,22 +1,44 @@
-# Release Notes v0.2.0
+# quickthink v0.2.1
 
-Date: 2026-02-18T18:43:38+00:00
+Released: 2026-08-04
 
-## Summary
-- Describe the user-visible improvements in this release.
+## What quickthink is for
 
-## Evidence Links
-- Eval report: unavailable
-- Compatibility snapshot: unavailable
+quickthink is a local-first CLI and Python library for shaping Ollama-backed LLM calls. It uses routing and a compact plan-then-answer scaffold for prompts that appear multi-step, while sending simpler prompts directly to the selected local model.
 
-## Compatibility Snapshot
-- No compatibility snapshot found.
+## What changed
 
-## Benchmark Notes
-Paste key metrics from the eval report here (win-rate, p50/p95 overhead, failure breakdown).
+- Prepared the first PyPI release path with a tag/version guard, distribution build, and package metadata validation before the protected trusted-publisher job can run.
+- Added clear install guidance for both the source-only and post-PyPI states.
+- Added `quickthink --version` for installed-package identification.
+- Synced package, citation, and Zenodo metadata to `0.2.1`.
 
-## Breaking Changes
-- None (or list explicitly).
+There are no inference-path changes in this release and no general model-quality improvement claim.
 
-## Upgrade Notes
-- Mention CLI/API changes and migration steps if any.
+## Install and first result
+
+After the PyPI publish job for this release has succeeded:
+
+```bash
+python -m pip install "quickthink==0.2.1"
+ollama pull qwen2.5:1.5b
+quickthink ask "Give me a 3-step plan to learn SQL basics" --model qwen2.5:1.5b
+```
+
+Before publication, install directly from the repository instead:
+
+```bash
+python -m pip install "quickthink @ git+https://github.com/hermes-labs-ai/quickthink.git"
+```
+
+## Requirements and limits
+
+- Python 3.9 or later, a running local [Ollama](https://ollama.com/) service, and a pulled model are required.
+- The supported model profiles are `qwen2.5:1.5b`, `mistral:7b`, and `gemma3:27b`; other models may run but are untuned.
+- quickthink is not a hosted API, training framework, or general agent-orchestration platform. It does not verify answer correctness; results remain model- and task-dependent.
+
+## Links
+
+- [Quickstart](https://github.com/hermes-labs-ai/quickthink#5-minute-quickstart)
+- [Known limitations](https://github.com/hermes-labs-ai/quickthink/blob/main/docs/KNOWN_LIMITATIONS.md)
+- [Release process](https://github.com/hermes-labs-ai/quickthink/blob/main/docs/release/RELEASE_PROCESS.md)

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "quickthink"
-version = "0.2.0"
+version = "0.2.1"
 description = "Compressed planning scaffold for local LLMs — latency-aware routing and structured output reliability"
 readme = "README.md"
 requires-python = ">=3.9"

src/quickthink/cli.py

@@ -2,6 +2,7 @@
 
 import json
 from datetime import datetime, timezone
+from importlib.metadata import version as package_version
 from pathlib import Path
 from typing import Optional
 
@@ -14,6 +15,25 @@
 app = typer.Typer(help="Compressed planning scaffold for local LLMs")
 
 
+def _version_callback(value: bool) -> None:
+    if value:
+        typer.echo(package_version("quickthink"))
+        raise typer.Exit()
+
+
+@app.callback()
+def main(
+    version: bool = typer.Option(
+        False,
+        "--version",
+        callback=_version_callback,
+        is_eager=True,
+        help="Show the installed quickthink version and exit.",
+    ),
+) -> None:
+    """Compressed planning scaffold for local LLMs."""
+
+
 @app.command()
 def list_models() -> None:
     for model, profile in MODEL_PROFILES.items():

tests/test_cli_version.py

@@ -0,0 +1,12 @@
+from importlib.metadata import version as package_version
+
+from typer.testing import CliRunner
+
+from quickthink.cli import app
+
+
+def test_version_reports_installed_package_version() -> None:
+    result = CliRunner().invoke(app, ["--version"])
+
+    assert result.exit_code == 0
+    assert result.output.strip() == package_version("quickthink")