textflowkit 0.1.13__tar.gz → 0.1.14__tar.gz

{textflowkit-0.1.13 → textflowkit-0.1.14}/AGENTS.md

@@ -10,7 +10,15 @@ The main assistant coordinates, plans and oversees. A worker invoked for an impl
 
 Follow Scott's active task selection; do not impose a project-wide Ollama default. His normal substantial-work/coordination preference is gpt-6.1-sol at medium, with gpt-6-luna at xhigh for small bounded work when available. Do not spawn a worker merely to switch models or add providers without authorization.
 
-For the explicitly authorized current Whistle release and issue #31 source-recovery work, Scott requested Ollama `deepseek-v4.1-flash:1017-cloud`, reasoning OFF. `scripts/run_deepseek_worker.py` is an explicit opt-in runner for those assigned DeepSeek units: it sends `think: false`, allows no model fallback, and records the actual response model. Ollama currently returns canonical ID `deepseek-v4.1-flash:1017` without the `-cloud` transport suffix. Do not confuse a worker's self-description with the transport receipts. Preserve existing DeepSeek integrations with reasoning off, but do not turn this task-specific instruction into the default for unrelated future work.
+## Required harness for all TextFlowKit DeepSeek coding (owner directive, 2026-10-19)
+
+Whenever DeepSeek is selected for coding work on TextFlowKit, use native **headless Codex exec through Ollama**, model **`deepseek-v4.1-flash:1017-cloud`**, thinking OFF (`model_reasoning_effort = "none"`). This is a standing project instruction, not limited to one release. The observed provider response ID is `deepseek-v4.1-flash:1017`; verify it from transport receipts, not model self-description.
+
+**Do not use `scripts/run_deepseek_worker.py`, another custom agent loop, or Claude Code for this work.** If the approved setup is missing, blocked or broken after restart, inspect/repair it within authorization or report the concrete blocker; never silently fall back, change the model, enable reasoning, or bypass safeguards.
+
+Use the dedicated worker worktree, separate CODEX_HOME/profile/temp directories, allowlisted environment and supported Windows workspace-write sandbox. No personal auth/settings/hooks/MCP configuration copied; no inherited provider/GitHub tokens; no nested workers or publishing. A worktree isolates changes, not all disk reads. Supply a scoped brief through stdin and retain native Codex events, provider model/reasoning evidence and actual tool results.
+
+Read `docs/deepseek-worker.md` before the next DeepSeek unit and after restart. Verified local launcher: `C:\Users\scott\Documents\Codex\2026-09-21\re\outputs\headless-codex-deepseek-20261019\run-codex.ps1`; local state and evidence are alongside it. This selects the harness whenever DeepSeek is used; it does not make DeepSeek the default for unrelated work.
 
 This does not change the application's transcription or AI-cleanup defaults. Whistle remains the default transcription engine.
 

{textflowkit-0.1.13 → textflowkit-0.1.14}/CHANGELOG.md

@@ -1,5 +1,14 @@
 # Changelog
 
+## v0.1.14 — 2026-10-19
+
+Documentation-only patch; core 0.1.14, unchanged optional fonts 0.1.6.
+
+- Persist the required native headless Codex workflow for all TextFlowKit DeepSeek coding: exact `deepseek-v4.1-flash:1017-cloud`, thinking off, isolated worktree/configuration/environment, supported Windows sandbox, and no silent fallback to the historical Python runner or Claude Code.
+- Link the workflow from contributor and integration documentation and update current package/manual/website release references.
+- No transcription, cleanup, streaming, dependency, or installer behavior change. The latest verified Windows installer remains v0.1.13 and retains its matching FFmpeg sources/notices. Native/live proof remains explicitly attributed to that earlier release; no new accuracy claim.
+- Math comparison fixtures, local logs, model credentials and machine-local launcher/profile files are not included.
+
 ## v0.1.13 — 2026-10-19
 
 Core: 0.1.13; unchanged optional fonts: 0.1.6. Whistle missing-speech fixes and a

{textflowkit-0.1.13 → textflowkit-0.1.14}/CONTRIBUTING.md

@@ -99,3 +99,7 @@ Read [LEGAL.md](LEGAL.md) before adding a source adapter.
 ## Security
 
 Do not open a public issue for a vulnerability. See [SECURITY.md](SECURITY.md).
+
+## DeepSeek coding workers
+
+When DeepSeek is selected for TextFlowKit coding, follow the required [isolated headless Codex workflow](docs/deepseek-worker.md): exact `deepseek-v4.1-flash:1017-cloud`, thinking off. Do not use the historical Python agent loop or Claude Code. This does not change application engines or make DeepSeek the default for unrelated work.

{textflowkit-0.1.13 → textflowkit-0.1.14}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: textflowkit
-Version: 0.1.13
+Version: 0.1.14
 Summary: Cross-platform media transcription toolkit: URL or file in, timestamped transcripts and subtitle files out.
 Project-URL: Homepage, https://www.textflowkit.org/
 Project-URL: Issues, https://github.com/scottconverse/textflowkit/issues
@@ -73,7 +73,7 @@ Cross-platform media transcription toolkit. **One core, one CLI, thin adapters.*
 [User manual](https://github.com/scottconverse/textflowkit/blob/main/docs/user-manual.md) ·
 [Developer and integration manual](https://github.com/scottconverse/textflowkit/blob/main/docs/adapters.md)
 
-**Current release: [v0.1.13](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13).**
+**Current release: [v0.1.14](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.14).**
 
 The [static site deployment](https://github.com/scottconverse/textflowkit/blob/main/docs/site-deployment.md) is hosted on Cloudflare
 Pages. GitHub remains the source and CI host; the website does not run the
@@ -134,7 +134,7 @@ DOCX and PDF are available through the optional `export` extra.
 
 ## Architecture
 
-![TextFlowKit shared-core architecture: five thin entry points feed one core. The CLI, Python API, MCP, HTTP API, and the local browser UI (a fifth door that mounts the existing HTTP app under /api and shares the same job store, adding transport only, not a new pipeline) sit above one core that acquires and decodes media, transcribes with the default Whistle engine or an explicitly selected openai-whisper, keeps job state and resume checkpoints in an optional SQLite store, adds optional speaker or translation postprocessing, and publishes TXT, SRT, VTT, JSON, Markdown, and optional DOCX/PDF exports.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.13/docs/assets/architecture-overview.svg)
+![TextFlowKit shared-core architecture: five thin entry points feed one core. The CLI, Python API, MCP, HTTP API, and the local browser UI (a fifth door that mounts the existing HTTP app under /api and shares the same job store, adding transport only, not a new pipeline) sit above one core that acquires and decodes media, transcribes with the default Whistle engine or an explicitly selected openai-whisper, keeps job state and resume checkpoints in an optional SQLite store, adds optional speaker or translation postprocessing, and publishes TXT, SRT, VTT, JSON, Markdown, and optional DOCX/PDF exports.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.14/docs/assets/architecture-overview.svg)
 
 The design principle is **one shared core, thin doors**. Everything of substance
 lives in the core; the interfaces are thin — the local browser UI is a fifth door
@@ -158,7 +158,7 @@ see [docs/architecture.md](https://github.com/scottconverse/textflowkit/blob/mai
 
 ## Live microphone streaming (opt-in)
 
-In v0.1.13, live streaming supports **Windows x86_64 only**. When
+In v0.1.14, live streaming supports **Windows x86_64 only**. When
 `TEXTFLOWKIT_STREAMING=1` is set, the developer HTTP app serves a `/stream`
 WebSocket and the local UI serves `/api/stream`, both transcribing a live microphone
 as you speak. A **runnable, self-hosted browser example** ships at
@@ -167,7 +167,7 @@ as you speak. A **runnable, self-hosted browser example** ships at
 No CDN, no analytics, no external fetch. Requires the `streaming` extra:
 
 ```bash
-python -m pip install "textflowkit[streaming]==0.1.13"
+python -m pip install "textflowkit[streaming]==0.1.14"
 ```
 
 It is **off by default** — with the switch unset the endpoint does not exist — and
@@ -233,7 +233,7 @@ installs `textflowkit-fonts` for offline multilingual PDF rendering:
 python -m pip install 'textflowkit[export]'
 ```
 
-## Windows desktop installer (v0.1.13)
+## Windows desktop installer (latest verified installer: v0.1.13)
 
 [Download Windows x64 setup](https://github.com/scottconverse/textflowkit/releases/download/v0.1.13/TextFlowKit-0.1.13-windows-x64-setup.exe) · [Matching sources and notices](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13)
 
@@ -362,7 +362,7 @@ textflowkit export ./transcript.json --format vtt
 needs no PyTorch, so a fresh `pip install textflowkit` stays small and pulls no
 torch; `openai-whisper` is an explicitly selectable opt-in engine.
 
-![Whistle bounded block run and durable resume: a decoded PCM WAV is split into 11-second cores with up to 2 seconds of context (each clip ≤ 15 s), each core runs as one owned child with forced telemetry-off flags, a partial checkpoint is written after each core, and a resume validates the source and decoded-WAV hashes then re-runs only the unfinished cores.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.13/docs/assets/whistle-resume.svg)
+![Whistle bounded block run and durable resume: a decoded PCM WAV is split into 11-second cores with up to 2 seconds of context (each clip ≤ 15 s), each core runs as one owned child with forced telemetry-off flags, a partial checkpoint is written after each core, and a resume validates the source and decoded-WAV hashes then re-runs only the unfinished cores.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.14/docs/assets/whistle-resume.svg)
 
 - **CPU only, no torch.** It runs a pinned native binary and downloads one pinned
   model on first use. Naming a GPU device is refused rather than silently
@@ -483,7 +483,7 @@ its own beyond the `http` extra. It ships with the core package as of v0.1.10:
 a regular install of the current release carries it.
 
 ```bash
-python -m pip install 'textflowkit[http,export]==0.1.13'
+python -m pip install 'textflowkit[http,export]==0.1.14'
 textflowkit-ui
 ```
 
@@ -554,7 +554,9 @@ browser interface share the job contract without blocking a request.
 
 ## Status
 
-**v0.1.13 release.** Fixes reproduced Whistle speech omissions. Native decode
+**v0.1.14 release.** Documentation-only patch: records the required isolated headless Codex workflow for DeepSeek coding contributors. Transcription, AI cleanup and streaming behavior are unchanged from v0.1.13; fonts remain 0.1.6. The latest verified Windows installer remains v0.1.13, with its matching source and notice bundles. No new native transcription proof is claimed for this documentation patch.
+
+**Whistle behavior introduced in v0.1.13.** Fixes reproduced Whistle speech omissions. Native decode
 inputs are now bounded to 15 seconds (11-second owned cores with up to 2 seconds
 of context per side), short interior speech gaps get a bounded second listen that
 accepts words only between matching timed anchors, and shared-context words are

{textflowkit-0.1.13 → textflowkit-0.1.14}/README.md

@@ -8,7 +8,7 @@ Cross-platform media transcription toolkit. **One core, one CLI, thin adapters.*
 [User manual](https://github.com/scottconverse/textflowkit/blob/main/docs/user-manual.md) ·
 [Developer and integration manual](https://github.com/scottconverse/textflowkit/blob/main/docs/adapters.md)
 
-**Current release: [v0.1.13](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13).**
+**Current release: [v0.1.14](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.14).**
 
 The [static site deployment](https://github.com/scottconverse/textflowkit/blob/main/docs/site-deployment.md) is hosted on Cloudflare
 Pages. GitHub remains the source and CI host; the website does not run the
@@ -69,7 +69,7 @@ DOCX and PDF are available through the optional `export` extra.
 
 ## Architecture
 
-![TextFlowKit shared-core architecture: five thin entry points feed one core. The CLI, Python API, MCP, HTTP API, and the local browser UI (a fifth door that mounts the existing HTTP app under /api and shares the same job store, adding transport only, not a new pipeline) sit above one core that acquires and decodes media, transcribes with the default Whistle engine or an explicitly selected openai-whisper, keeps job state and resume checkpoints in an optional SQLite store, adds optional speaker or translation postprocessing, and publishes TXT, SRT, VTT, JSON, Markdown, and optional DOCX/PDF exports.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.13/docs/assets/architecture-overview.svg)
+![TextFlowKit shared-core architecture: five thin entry points feed one core. The CLI, Python API, MCP, HTTP API, and the local browser UI (a fifth door that mounts the existing HTTP app under /api and shares the same job store, adding transport only, not a new pipeline) sit above one core that acquires and decodes media, transcribes with the default Whistle engine or an explicitly selected openai-whisper, keeps job state and resume checkpoints in an optional SQLite store, adds optional speaker or translation postprocessing, and publishes TXT, SRT, VTT, JSON, Markdown, and optional DOCX/PDF exports.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.14/docs/assets/architecture-overview.svg)
 
 The design principle is **one shared core, thin doors**. Everything of substance
 lives in the core; the interfaces are thin — the local browser UI is a fifth door
@@ -93,7 +93,7 @@ see [docs/architecture.md](https://github.com/scottconverse/textflowkit/blob/mai
 
 ## Live microphone streaming (opt-in)
 
-In v0.1.13, live streaming supports **Windows x86_64 only**. When
+In v0.1.14, live streaming supports **Windows x86_64 only**. When
 `TEXTFLOWKIT_STREAMING=1` is set, the developer HTTP app serves a `/stream`
 WebSocket and the local UI serves `/api/stream`, both transcribing a live microphone
 as you speak. A **runnable, self-hosted browser example** ships at
@@ -102,7 +102,7 @@ as you speak. A **runnable, self-hosted browser example** ships at
 No CDN, no analytics, no external fetch. Requires the `streaming` extra:
 
 ```bash
-python -m pip install "textflowkit[streaming]==0.1.13"
+python -m pip install "textflowkit[streaming]==0.1.14"
 ```
 
 It is **off by default** — with the switch unset the endpoint does not exist — and
@@ -168,7 +168,7 @@ installs `textflowkit-fonts` for offline multilingual PDF rendering:
 python -m pip install 'textflowkit[export]'
 ```
 
-## Windows desktop installer (v0.1.13)
+## Windows desktop installer (latest verified installer: v0.1.13)
 
 [Download Windows x64 setup](https://github.com/scottconverse/textflowkit/releases/download/v0.1.13/TextFlowKit-0.1.13-windows-x64-setup.exe) · [Matching sources and notices](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13)
 
@@ -297,7 +297,7 @@ textflowkit export ./transcript.json --format vtt
 needs no PyTorch, so a fresh `pip install textflowkit` stays small and pulls no
 torch; `openai-whisper` is an explicitly selectable opt-in engine.
 
-![Whistle bounded block run and durable resume: a decoded PCM WAV is split into 11-second cores with up to 2 seconds of context (each clip ≤ 15 s), each core runs as one owned child with forced telemetry-off flags, a partial checkpoint is written after each core, and a resume validates the source and decoded-WAV hashes then re-runs only the unfinished cores.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.13/docs/assets/whistle-resume.svg)
+![Whistle bounded block run and durable resume: a decoded PCM WAV is split into 11-second cores with up to 2 seconds of context (each clip ≤ 15 s), each core runs as one owned child with forced telemetry-off flags, a partial checkpoint is written after each core, and a resume validates the source and decoded-WAV hashes then re-runs only the unfinished cores.](https://raw.githubusercontent.com/scottconverse/textflowkit/v0.1.14/docs/assets/whistle-resume.svg)
 
 - **CPU only, no torch.** It runs a pinned native binary and downloads one pinned
   model on first use. Naming a GPU device is refused rather than silently
@@ -418,7 +418,7 @@ its own beyond the `http` extra. It ships with the core package as of v0.1.10:
 a regular install of the current release carries it.
 
 ```bash
-python -m pip install 'textflowkit[http,export]==0.1.13'
+python -m pip install 'textflowkit[http,export]==0.1.14'
 textflowkit-ui
 ```
 
@@ -489,7 +489,9 @@ browser interface share the job contract without blocking a request.
 
 ## Status
 
-**v0.1.13 release.** Fixes reproduced Whistle speech omissions. Native decode
+**v0.1.14 release.** Documentation-only patch: records the required isolated headless Codex workflow for DeepSeek coding contributors. Transcription, AI cleanup and streaming behavior are unchanged from v0.1.13; fonts remain 0.1.6. The latest verified Windows installer remains v0.1.13, with its matching source and notice bundles. No new native transcription proof is claimed for this documentation patch.
+
+**Whistle behavior introduced in v0.1.13.** Fixes reproduced Whistle speech omissions. Native decode
 inputs are now bounded to 15 seconds (11-second owned cores with up to 2 seconds
 of context per side), short interior speech gaps get a bounded second listen that
 accepts words only between matching timed anchors, and shared-context words are

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/adapters.md

@@ -1,6 +1,10 @@
 # Developer and integration manual
 
-Current release: [v0.1.13](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13).
+Current release: [v0.1.14](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.14).
+
+v0.1.14 updates developer workflow documentation only; runtime behavior is unchanged. The latest verified Windows installer remains v0.1.13 with its matching sources and notices.
+For DeepSeek coding contributions, use the required [headless Codex worker workflow](deepseek-worker.md); this is a development tool, not a transcription-engine setting.
+
 Start with the [user manual](user-manual.md) for everyday use. Harness versions
 in the compatibility table below identify **historical test sessions**, not
 the version of TextFlowKit currently published.
@@ -553,7 +557,7 @@ observation rather than forwarding the caller's header, and keep
 
 ## Live microphone streaming (`/stream`, opt-in)
 
-In v0.1.13, native live streaming (introduced in v0.1.11) supports **Windows x86_64 only**. Other live
+In v0.1.14, native live streaming (introduced in v0.1.11) supports **Windows x86_64 only**. Other live
 platforms are unsupported; existing standalone defaults and platform support are
 unchanged. Physical browser microphone capture was not verified. The public site
 is static documentation and serves no streams. Both local server surfaces can
@@ -564,7 +568,7 @@ exist and a connection is refused before any session is allocated - and it needs
 `streaming` extra (`websockets`).
 
 ```powershell
-python -m pip install "textflowkit[streaming]==0.1.13"
+python -m pip install "textflowkit[streaming]==0.1.14"
 # Native Windows PowerShell:
 $env:TEXTFLOWKIT_API_TOKEN="choose-a-long-random-token"
 $env:TEXTFLOWKIT_STREAMING="1"
@@ -604,7 +608,7 @@ supplies FastAPI and uvicorn, and adds no new dependency). It has no public site
 of its own: it is a local app, not a hosted service.
 
 ```bash
-python -m pip install 'textflowkit[http,export]==0.1.13'
+python -m pip install 'textflowkit[http,export]==0.1.14'
 textflowkit-ui                 # serve on 127.0.0.1 and open a browser
 textflowkit-ui --no-browser --port 8901
 ```

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/architecture.md

@@ -1,4 +1,4 @@
-# Architecture and job lifecycle — v0.1.13
+# Architecture and job lifecycle — v0.1.14
 
 This guide describes how textflowkit is put together and how a job moves through
 it. It is orientation for maintainers and integrators; it is not a deployment

textflowkit-0.1.14/docs/deepseek-worker.md

@@ -0,0 +1,29 @@
+# Required DeepSeek coding harness
+
+Owner directive, 2026-10-19: all DeepSeek coding work on TextFlowKit uses native headless Codex exec through Ollama. No custom agent loop or Claude Code. This remains binding after restart unless Scott changes it.
+
+- Exact requested model: `deepseek-v4.1-flash:1017-cloud`. Provider-reported canonical ID: `deepseek-v4.1-flash:1017`.
+- Thinking off: `model_reasoning_effort = "none"`; verify actual provider effort and reasoning-token receipts.
+- No model/provider/harness fallback. Missing setup is a repair task or blocker, not permission to revert to `scripts/run_deepseek_worker.py`.
+- Use dedicated worktree/branch, isolated CODEX_HOME/profile/temp, cleared-and-allowlisted process environment, no copied personal auth/settings/hooks/MCP configuration, and native Windows workspace-write sandbox. Do not bypass enforced controls. This is not a VM or full read-confidentiality boundary.
+- Send a scoped brief on stdin. Preserve native JSONL events, stderr, session/metadata, PID ownership and terminal state. Workers do not publish or launch nested workers. Main assistant coordinates/audits; worker implements the assigned unit.
+
+## Local installation on Scott's Windows machine
+
+- Worktree: `C:\Users\scott\Desktop\Code\textflowkit-headless-worker`
+- Branch: `worker/headless-codex-deepseek`
+- Setup directory: `C:\Users\scott\Documents\Codex\2026-09-21\re\outputs\headless-codex-deepseek-20261019`
+- Launcher: `run-codex.ps1` in that directory (Brief, Unit and optional Resume parameters).
+- Configuration: `codex-home\config.toml`; model metadata: `codex-home\models.json`.
+- Endpoint: `http://127.0.0.1:11434/v1`. IPv6 localhost refused the setup probe; no service configuration was changed.
+- Tested Codex: `0.163.0-alpha.10.2`; launcher pins the tested executable path. If an app upgrade removes that path, verify the replacement executable/configuration and affected behavior, preserving the exact model and thinking-off setting. Do not silently revert to another harness.
+- Evidence: `verification.json`, `runs\`, `README.md`, current `STATE.md` alongside the launcher.
+
+## Restart checklist
+
+1. Read repository AGENTS.md, this document and the setup STATE.md.
+2. Inspect worktree status and the scoped brief; preserve unrelated changes.
+3. Check the launcher/config still select exact model, none effort, isolation and sandbox. Reuse unaffected verification evidence; retest changed inputs or unresolved acceptance only.
+4. Use native Codex, never the historical Python runner. Report actual provider/model, reasoning setting and tool results separately from configuration.
+
+The old Python runner remains historical repository code; its presence is not authorization to use it. This instruction does not change Whistle, application cleanup models, normal non-DeepSeek preferences, or existing release authorization boundaries.

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/index.html

@@ -206,7 +206,7 @@
   <main id="main">
     <div class="wrap hero">
       <div>
-        <span class="eyebrow">Open source · Apache 2.0 · v0.1.13</span>
+        <span class="eyebrow">Open source · Apache 2.0 · v0.1.14</span>
         <h1>Media in.<br><em>Useful text</em> out.</h1>
         <p class="lead">Transcribe video and audio into timestamped, searchable text. One shared core for your command line, AI harness, and application.</p>
         <div class="actions">
@@ -214,7 +214,8 @@
           <a class="button secondary" href="https://pypi.org/project/textflowkit/">Install from PyPI <span aria-hidden="true">&nbsp;↗</span></a>
         </div>
         <p class="micro">Self-hosted toolkit · Native Windows, macOS, Linux · No account required</p>
-        <p class="micro">Latest verified release: <a href="https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13">v0.1.13</a> · <a href="https://github.com/scottconverse/textflowkit/blob/main/docs/user-manual.md">User manual</a></p>
+        <p class="micro">Latest verified release: <a href="https://github.com/scottconverse/textflowkit/releases/tag/v0.1.14">v0.1.14</a> · <a href="https://github.com/scottconverse/textflowkit/blob/main/docs/user-manual.md">User manual</a></p>
+        <p class="micro">v0.1.14 is a developer-documentation patch; transcription behavior is unchanged. The verified Windows installer remains v0.1.13.</p>
       </div>
       <div class="terminal" aria-label="Example transcription command and output">
         <div class="terminal-top"><span class="dot"></span><span class="dot"></span><span class="dot"></span><span class="terminal-label">textflowkit / terminal</span></div>
@@ -308,7 +309,7 @@
       <div class="install-grid">
         <div>
           <ul>
-            <li><strong>Install and launch:</strong> <code>python -m pip install "textflowkit[http,export]==0.1.13"</code> then <code>textflowkit-ui</code>. It adds no dependency beyond the <code>http</code> extra.</li>
+            <li><strong>Install and launch:</strong> <code>python -m pip install "textflowkit[http,export]==0.1.14"</code> then <code>textflowkit-ui</code>. It adds no dependency beyond the <code>http</code> extra.</li>
             <li><strong>Desktop shortcut (Windows):</strong> <code>textflowkit-ui --create-shortcut</code> makes a Start Menu entry that launches with <code>pythonw</code> (no console) and needs no <code>pywin32</code>. That shortcut is the normal way to open it.</li>
             <li><strong>Loopback only:</strong> it binds <code>127.0.0.1</code> and is never reachable off the machine, whatever <code>TEXTFLOWKIT_ALLOW_REMOTE</code> is set to.</li>
             <li><strong>Durable jobs, cancel, and resume:</strong> jobs and checkpoints are stored durably; you can cancel a running job and resume a saved one. Shutdown <em>drains</em> — it waits for current jobs to finish rather than cancelling them.</li>
@@ -319,7 +320,7 @@
           <a class="button secondary" href="https://github.com/scottconverse/textflowkit/blob/main/docs/user-manual.md#7-local-browser-interface">Read the local UI guide ↗</a>
         </div>
         <div>
-          <pre class="codebox"><span class="prompt">$ </span>python -m pip install "textflowkit[http,export]==0.1.13"
+          <pre class="codebox"><span class="prompt">$ </span>python -m pip install "textflowkit[http,export]==0.1.14"
 <span class="prompt">$ </span>textflowkit-ui --create-shortcut
 <span class="comment"># then open it from the Start Menu shortcut</span></pre>
           <p class="note"><strong>Public page vs. local app.</strong> This landing page is the public, static site. The browser workspace is a separate, local program: it never runs here and is not reachable from the internet. v0.1.12 added optional UI-only local Ollama cleanup and native Windows packaging (that cleanup remains as introduced in v0.1.12); existing standalone defaults, file/URL jobs, CLI, and MCP behavior remain unchanged. This public site serves no streaming endpoint.</p>
@@ -387,7 +388,7 @@
           <li>Find the right setup for AMD ROCm in the install guide before installing dependencies — this applies to the optional <code>openai-whisper</code> engine and to diarization.</li>
         </ul>
         <a class="button secondary" href="https://github.com/scottconverse/textflowkit/blob/main/docs/install.md">Read install guide ↗</a>
-        <p class="note">New here? Follow the <a href="https://github.com/scottconverse/textflowkit/blob/main/docs/user-manual.md">v0.1.13 user manual</a> for CLI, MCP, HTTP, the local browser interface, export, and troubleshooting examples, or the <a href="https://github.com/scottconverse/textflowkit/blob/main/docs/adapters.md">developer and integration manual</a> for harness configuration and production settings.</p>
+        <p class="note">New here? Follow the <a href="https://github.com/scottconverse/textflowkit/blob/main/docs/user-manual.md">v0.1.14 user manual</a> for CLI, MCP, HTTP, the local browser interface, export, and troubleshooting examples, or the <a href="https://github.com/scottconverse/textflowkit/blob/main/docs/adapters.md">developer and integration manual</a> for harness configuration and production settings.</p>
       </div>
       <div>
         <pre class="codebox"><span class="prompt">$ </span>python -m pip install "textflowkit[mcp,http]"

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/install.md

@@ -1,6 +1,8 @@
 # Install notes
 
-Current release: [v0.1.13](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13).
+Current release: [v0.1.14](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.14).
+
+v0.1.14 updates developer workflow documentation only; runtime behavior is unchanged. The latest verified Windows installer remains v0.1.13 with its matching sources and notices.
 Use `textflowkit --version` to confirm the installed version. For everyday
 commands and outputs, start with the [user manual](user-manual.md).
 
@@ -55,7 +57,7 @@ The local browser workspace **ships in the core package as of v0.1.10**. A
 regular install of the current release carries the `textflowkit-ui` command:
 
 ```bash
-python -m pip install 'textflowkit[http,export]==0.1.13'
+python -m pip install 'textflowkit[http,export]==0.1.14'
 textflowkit-ui
 ```
 
@@ -84,14 +86,14 @@ them; cancel a job first to stop it early. See the
 
 ## Live microphone streaming
 
-Live streaming in v0.1.13 (introduced in v0.1.11) is **Windows x86_64 only, opt-in and off by default**.
+Live streaming in v0.1.14 (introduced in v0.1.11) is **Windows x86_64 only, opt-in and off by default**.
 Other live platforms are unsupported; existing standalone platform support and
 Whistle defaults are unchanged. Physical browser microphone capture has not been
 verified; the public website is static documentation and does not accept streams. It needs one extra dependency
 (the WebSocket implementation for the ASGI server):
 
 ```bash
-python -m pip install "textflowkit[streaming]==0.1.13"
+python -m pip install "textflowkit[streaming]==0.1.14"
 ```
 
 Set `TEXTFLOWKIT_STREAMING=1` to serve it, then open the shipped browser example.

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/release-checklist.md

@@ -1,6 +1,12 @@
 # Release verification checklist
 
-## v0.1.13 release preparation
+## v0.1.14 documentation-only patch
+
+Core 0.1.14; fonts 0.1.6 reused unchanged. This release changes developer workflow documentation and current version metadata only, not application logic/dependencies. Run `python scripts/verify_release_versions.py --tag v0.1.14`, `python scripts/verify_readme_release.py --version v0.1.14`, affected release-surface tests and required CI. Merge only after green PR CI; require exact merged-main CI before tagging/publishing. Verify published distribution hashes, fresh installed version and the deployed website.
+
+The verified Windows installer and its corresponding sources/notices remain v0.1.13; keep those download links explicitly versioned. Do not relabel/rebuild unchanged installer artifacts merely for this documentation patch. Reuse unaffected v0.1.13 native/live evidence with its original commit/version and do not claim a new live YouTube or native accuracy verification for v0.1.14. No changes to the live-check script or its requirements.
+
+## Historical v0.1.13 release preparation
 
 Core is 0.1.13; fonts remain 0.1.6, unchanged. Run the existing guards from the candidate root:
 

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/roadmap.md

@@ -262,3 +262,8 @@ the full five-role audit of v0.1.6 main commit
 Deferral does not remove these findings or waive their acceptance criteria.
 Core engineering/runtime bugs and the major documentation corrections remain
 in the active repair goal. See the [changelog](../CHANGELOG.md).
+
+## v0.1.14 — developer workflow documentation
+
+- [x] Ship v0.1.14 review follow-ups: persist the headless-Codex-only DeepSeek coding policy across restarts; require the exact 1017-cloud model, thinking off, isolated worktree/config/environment and native Windows sandbox, with no silent fallback.
+- [x] Update current package/documentation release references; retain the verified v0.1.13 Windows installer and historical native evidence unchanged. No new application capability.

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/streaming.md

@@ -1,4 +1,4 @@
-# Live microphone streaming — v0.1.13
+# Live microphone streaming — v0.1.14
 
 TextFlowKit can transcribe a **live microphone** as you speak: audio is captured in
 the browser, resampled to 16 kHz mono, and streamed to a local WebSocket, and the
@@ -9,7 +9,7 @@ commands, and the limits.
 The capability is **opt-in, loopback-only, and verified on this machine's native
 engine for bounded live runs** — 65 s and 300 s end-to-end native streams ran
 locally on Windows x86_64 (see *Limits and known constraints*). Native live support is **Windows x86_64 only**; other live platforms are
-unsupported. This v0.1.13 release surface (introduced in v0.1.11) is gated behind an environment switch
+unsupported. This v0.1.14 release surface (introduced in v0.1.11) is gated behind an environment switch
 so a default install never serves it. The public website is static documentation,
 not a streaming service. Publication and remote CI are separate release gates.
 
@@ -45,7 +45,7 @@ audio buffer is allocated.
 ### Option A — the local browser UI (point and click)
 
 ```bat
-python -m pip install "textflowkit[streaming]==0.1.13"
+python -m pip install "textflowkit[streaming]==0.1.14"
 set TEXTFLOWKIT_STREAMING=1
 textflowkit-ui
 ```
@@ -61,7 +61,7 @@ first WebSocket message; it is never written into the page, its JavaScript, or a
 ### Option B — the developer HTTP API (programmatic clients)
 
 ```bat
-python -m pip install "textflowkit[streaming]==0.1.13"
+python -m pip install "textflowkit[streaming]==0.1.14"
 set TEXTFLOWKIT_STREAMING=1
 set TEXTFLOWKIT_API_TOKEN=choose-a-long-random-token
 textflowkit-http --streaming

{textflowkit-0.1.13 → textflowkit-0.1.14}/docs/user-manual.md

@@ -1,4 +1,6 @@
-# TextFlowKit user manual — v0.1.13
+# TextFlowKit user manual — v0.1.14
+
+v0.1.14 is a documentation-only patch. Transcription behavior is unchanged from v0.1.13; the latest verified Windows installer remains v0.1.13. Developer workflow changes are described in the [integration manual](adapters.md).
 
 TextFlowKit turns a local audio/video file or a supported media URL into a
 timestamped transcript. It is a **self-hosted developer tool**, not a hosted
@@ -79,7 +81,7 @@ Install Python 3.10 or later and `ffmpeg`/`ffprobe` on `PATH`. For a standard
 CPU setup, install the current release from PyPI:
 
 ```bash
-python -m pip install 'textflowkit[export,mcp,http]==0.1.13'
+python -m pip install 'textflowkit[export,mcp,http]==0.1.14'
 textflowkit --version
 textflowkit doctor
 textflowkit selftest
@@ -347,7 +349,7 @@ point from the developer HTTP API, reuses the same pipeline and job store, and i
 served on loopback only.
 
 ```bash
-python -m pip install 'textflowkit[http,export]==0.1.13'
+python -m pip install 'textflowkit[http,export]==0.1.14'
 textflowkit-ui
 ```
 
@@ -439,7 +441,7 @@ and nowhere else.
 
 It is **off by default**: with the switch unset the endpoint
 does not exist. It needs the `streaming` extra
-(`python -m pip install "textflowkit[streaming]==0.1.13"`). This is a capability of the
+(`python -m pip install "textflowkit[streaming]==0.1.14"`). This is a capability of the
 **servers**, not the command-line `textflowkit` tool — the CLI never opens a
 microphone. See [streaming.md](streaming.md) for the full contract and the
 programmatic API.
@@ -523,8 +525,8 @@ passes with Whistle's own real speech output.
 
 ## 9. Release and help
 
-- [v0.1.13 GitHub release](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.13)
-- [Core package 0.1.13 on PyPI](https://pypi.org/project/textflowkit/0.1.13/) and [unchanged optional font package 0.1.6](https://pypi.org/project/textflowkit-fonts/0.1.6/)
+- [v0.1.14 GitHub release](https://github.com/scottconverse/textflowkit/releases/tag/v0.1.14)
+- [Core package 0.1.14 on PyPI](https://pypi.org/project/textflowkit/0.1.14/) and [unchanged optional font package 0.1.6](https://pypi.org/project/textflowkit-fonts/0.1.6/)
 - [Release verification procedure](release-checklist.md), [security policy](../SECURITY.md), and [issues](https://github.com/scottconverse/textflowkit/issues)
 
 The v0.1.13 release fixes reproduced Whistle speech omissions with a 15-second

{textflowkit-0.1.13 → textflowkit-0.1.14}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "textflowkit"
-version = "0.1.13"
+version = "0.1.14"
 description = "Cross-platform media transcription toolkit: URL or file in, timestamped transcripts and subtitle files out."
 readme = "README.md"
 requires-python = ">=3.10"

{textflowkit-0.1.13 → textflowkit-0.1.14}/src/textflowkit/__init__.py

@@ -18,7 +18,7 @@ from textflowkit.core.streaming_events import (
     WorkerProcessError,
 )
 
-__version__ = "0.1.13"
+__version__ = "0.1.14"
 
 __all__ = [
     "PipelineError",