v1.1.8

Added

• Bare-command welcome detects missing ffmpeg and ffprobe alongside chromium. Lists per-OS install hints (apt / pacman / dnf / brew / winget).
• Init wizard prints the resolved cwd and gates further prompts behind an explicit confirm before scaffolding.

Docs

• README first-demo section leads with cd <product-root> and explains why running init from a non-product directory breaks the scaffold's defaults.
• New "After installing: run the doctor first" section pointing at the system-only showrunner doctor pass.

v1.1.7

Fixed

• understand --agent was spawning claude with cwd = process.cwd(), which is typically the scaffold directory rather than the product codebase. The agent had nothing useful to read and timed out.

Added

• --project-dir <path> flag on showrunner understand for explicit per-invocation override.
• project.codebase_root field on the config schema, resolved relative to the config's directory.
• Resolution order when --agent is set: --project-dir flag, then project.codebase_root, then configDir/.. (when -c is given), then process.cwd(). The resolved path is logged alongside the rule that selected it.
• Scaffolded demo.yaml writes codebase_root: .. by default.

Changed

• Claude spawn passes --allowedTools "Read,Glob,Grep" explicitly.
• Agent timeout reduced from 180s to 90s.

v1.1.6

Added

• showrunner understand --agent delegates project exploration to the local claude CLI. Spawns claude -p --output-format json; the agent uses its native Read/Glob/Grep tools to walk the project and synthesise product_model.json matching the existing Zod schema. Mutually exclusive with --interactive. Errors if claude is missing from PATH.

Changed

• Error message for empty comprehension.sources now names all three options: add a source, --interactive, or --agent.

Known limitations

• type: codebase entries in comprehension.sources remain non-functional in the non---agent path. Directories are silently skipped and include / exclude globs are ignored. Wiring real codebase support to the document path remains future work.

v1.1.5

Added

• showrunner set-target -c <config> --url <url> re-probes a URL and rewrites recording.target_url in demo.yaml. --force skips the probe.
• Port-scan during init. When the typed URL doesn't respond, probes common dev ports on localhost in parallel (3000, 3001, 4321, 5173, 5174, 8000, 8080) and offers responding ports as a select.
• Agent-driven dev-server discovery. When the LLM provider is agent_bridge and claude is on PATH, the wizard offers to inspect the project, propose a dev-server command + URL, spawn it detached and unref'd, then poll the URL for up to 60 seconds. Reports the spawned PID.

Changed

• init's URL section is now a cascade: direct probe → port-scan → optional agent discovery → fallback. Each step short-circuits on success.
• src/setup/targetProbe.ts is shared between init and set-target.

Known limitations

• Spawned dev-server PIDs are not tracked across runs. Stop manually with kill <pid>.
• set-target rewrites demo.yaml via yaml.dump, which strips comments.
• Agent discovery only fires when the LLM provider is agent_bridge.

v1.1.4

Added

• showrunner init is interactive by default when stdout is a TTY. The wizard detects the environment (claude CLI, ffmpeg/ffprobe, chromium, provider env vars), then prompts through project name, resolution, LLM provider (defaults to agent_bridge when claude is detected), inline API key paste, TTS provider, inline API key paste, target URL with a HEAD probe.
• A real .env is written with the pasted keys alongside .env.example.
• --yes flag skips prompts and uses defaults or passed flags. Automatically forced when stdout is not a TTY.
• @clack/prompts dependency for the wizard UI.

Changed

• init next-steps footer adapts to wizard state. Skips the "edit .env" step when keys were pasted; lists only the still-missing keys otherwise.

v1.1.3

Changed

• Bare-command welcome rewritten. Each state (chromium missing / no project / in project) now shows one next command and one line of context, instead of a 15-line block that duplicated init's footer.
• init next-steps footer split conflated steps. Each required env var is listed individually with the provider's dashboard URL; the agent_bridge alternative is a single parenthetical at the end.

v1.1.2

Fixed

• showrunner --version crashed on fresh global install with ERR_MODULE_NOT_FOUND for @babel/parser. The babel runtime dependencies (@babel/parser, @babel/traverse, @babel/types) used by the instrument stage were in devDependencies and therefore absent from the published tarball. Moved to dependencies. @types/babel__traverse stays in devDependencies (type-only).

Added

• showrunner install-browser subcommand. Wraps the bundled playwright-core CLI directly, sidestepping the "install your dependencies first" warning that bare npx playwright install emits when invoked outside a Node project.
• Welcome detects missing chromium and surfaces an install-browser hint.

v1.1.1

Changed

• Replace playwright with playwright-core. Eliminates the ~300 MB browser auto-download on npm install. npx playwright install chromium is now a required post-install step (enforced by doctor).
• doctor failure rows include per-OS install hints for ffmpeg and ffprobe (apt / pacman / dnf on Linux, brew on macOS, winget / choco on Windows).
• doctor probes the root / admin Playwright cache and warns when a browser is present there but missing from the user cache (the typical aftermath of sudo npx playwright install).

Removed

• fluent-ffmpeg and @types/fluent-ffmpeg. The package was listed but never imported; ffmpeg invocations have always gone through child_process.spawn via the internal runFfmpeg() wrapper.

Fixed

• Generated Playwright spec at src/manifest/playwrightCodegen.ts imported expect from 'playwright', which doesn't export it. The assert_visible action now emits locator.waitFor({ state: 'visible' }).

Docs

• README leads with npx @kadj-amoah/showrunner. Browser bootstrap (npx playwright install chromium) is called out as required.
• Troubleshooting note added for Arch / CachyOS / Fedora users who see Playwright's "OS not officially supported" warning during browser install.

v1.1.0 — first release

Changelog

All notable changes to Showrunner are documented here. Format follows Keep a Changelog; the project tracks loose semver — minor bumps for new capability, patch for fixes.

1.1.0 — 2026-05-23

First tagged release. Versioned 1.1.0 rather than 0.1.0 because the project has been tracked as "v1.1 — reliability hardening + provider-agnostic refactor" throughout development; this is the first cut where the pipeline works end-to-end against real Next.js targets and the LLM + TTS layers are swappable. No 1.0.0 was ever published.

Added

• Provider-agnostic LLM layer (llm.default + per-stage llm.overrides). Four providers:
  • anthropic — Claude via the official SDK with structured outputs.
  • openai — gpt-4o-style models via response_format: json_schema with json_object fallback.
  • agent_bridge — spawn a headless CLI agent like claude -p --output-format json; no API key on file required. Supports spawn and file_poll modes.
  • custom — dynamic-import an operator-supplied module conforming to the LLMProvider interface.
• Provider-agnostic TTS layer (voiceover.provider). Three providers:
  • elevenlabs — the only one that returns per-character alignment.
  • openai — tts-1-hd via the audio API.
  • custom — dynamic-import for in-house TTS pipelines.
• alignment_strategy: required | best_effort — when a TTS provider doesn't return alignment, switch automatically to per-segment synthesis (one TTS call per segment) so audio doesn't get sliced across mid-word boundaries.
• showrunner doctor — preflight checks before any run. Validates config syntax, provider env vars, ffmpeg + ffprobe on PATH, Playwright chromium binary, target URL reachability, free disk on every output dir, free memory + computed ffmpeg thread cap, and lifecycle script executability. Wired into run implicitly (escape with --skip-doctor).
• showrunner record-actions with scroll capture. Drives a headed browser; captures click / input / change / keydown / submit / scroll events; coalesces them into a clean action stream. Scroll capture is 250 ms debounced with direction coalescing.
• DOM preflight in the script stage — scrapes the live target's actionable selector inventory (≈60–80 entries on a typical marketing site) before asking the LLM for a manifest. The LLM is constrained to inventory selectors only, with a one-shot remediation retry if it strays.
• Resource preflight in record and mux stages — checks free disk against estimated artifact sizes, computes an ffmpeg -threads cap from free RAM × resolution × quality preset (SHOWRUNNER_FFMPEG_THREADS env override available).
• Structured ffmpeg errors — categorized as oom | no_space | codec_missing | permission | unknown, with per-category remediation hints.
• --resolution flag on both init and run. Presets: low (854×480), standard (720p, the default), high (1080p), extreme (4K). On run it overrides both recording.viewport and output.resolution in one step so they stay matched and mux doesn't upscale.
• init flags: --llm-provider, --tts-provider, --resolution. Scaffolds a demo.yaml with the right provider blocks, an .env.example listing only the env vars that combination actually needs (deduplicated when LLM + TTS both use OpenAI), and a real docs/PRD.md stub with section guides.
• Legacy config auto-migration — pre-v0.1 demo.yaml files with flat voiceover.{voice_id, model, ...} fields and no llm block are normalized at load time. Existing configs keep working with zero edits.
• validate --strict — exit nonzero on missing provider env vars (otherwise warns).
• Captions — SRT + VTT emitted alongside the final MP4 when output.captions.enabled is set. Whole-segment cues are used as a fallback when per-character alignment isn't available.

Changed

• engines.node bumped to >=20.6 (needed for process.loadEnvFile).
• init defaults: script.vo_review_gate: false (was true), recording.viewport: 1280×720 (was 1920×1080), output.resolution: 1280x720. Each is opt-in to the larger/stricter value via a one-line edit.
• tsup configured with skipNodeModulesBundle: true and target: 'node20'. Built dist/cli.js dropped from 1.75 MB to 247 KB and no longer crashes with Dynamic require of "tty" is not supported — npm link and npm i -g are viable now.
• README rewritten for v0.1 — install path (with npm link), full prereqs (Node ≥ 20.6, ffmpeg, chromium), provider-choice tables, pipeline-per-stage table, daily-commands cookbook, troubleshooting section.

Fixed

• understand --interactive no longer silently exits after the second question on piped stdin. Replaced readline/promises.question() with a queued 'line'-event reader that handles both TTY and pipe semantics.
• understand --output <path> now correctly takes precedence over project.product_model in the config (was silently overridden).
• Script-stage prompt now bans Tailwind arbitrary-value class names (.max-w-[1280px]) which aren't legal CSS and crashed Playwright with a SyntaxError. A post-validator catches violations and retries once with a remediation prompt before accepting the manifest.
• looksUnstable() heuristic now catches Next.js compiled-CSS classes like __variable_3eb911 (leading underscores were previously slipping past).
• mux output write now falls back to a timestamped sibling MP4 when the canonical path is locked by a media player, with a warning, instead of failing the run.

Known limitations (Tier 3 — coded but not exercised end-to-end)

• OpenAI LLM + OpenAI TTS provider paths are wired but were not run with a real API key in v0.1 validation. The agent_bridge + ElevenLabs combination is the validated default.
• agent_bridge file_poll mode is coded but only spawn mode was exercised.
• Custom provider modules (LLM + TTS) — dynamic-import loader is wired but no reference implementation has been swapped in.
• instrument, capture-auth, trace, preview, rerun-segment, print-vo, approve-vo commands are mostly thin wrappers but were not stress-tested in v0.1.
• Background music mix and title-card logo / custom-font rendering are coded but were not exercised.
• 1920×1080 mux remains RAM-sensitive on tight boxes; the thread-cap heuristic mitigates but the safe default is --resolution standard.

Validation artifact

A real end-to-end run against the Credstone Atlas marketing site produced output/ct-website-atlas_demo.mp4 (4.3 MB, H.264 + AAC, 1280×720, 48.2 s, with SRT + VTT) using:

• agent_bridge LLM (via local claude CLI, no Anthropic key on file)
• ElevenLabs TTS with alignment_strategy: required
• DOM preflight + selector validator producing inventory-only manifest selectors
• Full mux preflight + thread cap on a 5.9 GB RAM machine with 1.1 GB free at run time