bitterclip-marketing
sheetgenius/bitterclip-marketing
BitterClip is a speaker-aware media workspace for turning long recordings into source-linked clips.
Instalação
npx convez-ai sheetgenius/bitterclip-marketing1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
README
A tradução para o português deste README ainda está na fila.
BitterClip
BitterClip is a speaker-aware media workspace for turning long recordings into source-linked clips. It is designed for AI agents that can search transcript context and for human operators who need to verify the final cut before publishing.
It is built for podcasts, interviews, founder calls, demos, livestreams, expert conversations, and recurring shows where the strongest moments are buried inside long-form media. BitterClip keeps the recording, transcript, speaker identity, clip boundary, export, publishing state, and source history connected.
Place In Bitter
Bitter is an agent-operable software environment. It gives AI agents persistent workspaces where repositories, credentials, hosting, email, tests, deploys, logs, checks, and work history are connected.
The Bitter constellation is the set of public sites, products, applications, and supporting repositories that share that operating model:
- agents work against real source material
- humans review, steer, and approve the work
- deployed systems retain enough context to keep improving
- outputs stay connected to their source history and verification path
BitterClip applies that model to recorded media. A long recording becomes structured source material. A transcript becomes an agent-operable workspace. A clip remains linked to who said what, where it happened, what came before and after, how it was exported, and what has already been clipped.
Product Model
BitterClip is not a generic AI shorts generator. It is a media workbench built around source-linked verification.
The core object chain is:
Recording -> Transcript -> Speakers -> Moments -> Clips -> Exports -> Publishing
The main workflow is:
- Upload a recording.
- Build a time-aligned transcript.
- Identify and confirm speakers.
- Use an AI agent to identify candidate Moments.
- Open the Moment against the real media and surrounding transcript.
- Verify the speaker, context, and boundaries.
- Save the approved Moment as a Clip and create Exports.
- Use Publishing connections for approved Clips when available.
- Preserve source links and project history for future recordings.
This matters because useful clips are rarely just isolated quotes. They depend on speaker identity, setup, response, payoff, and the surrounding context that keeps the cut honest.
Repository Role
This repository contains the public BitterClip website at bitterclip.com. It is the public semantic entry point for BitterClip inside the Bitter constellation.
It is not the private BitterClip product application, Rails workspace, media-processing backend, billing system, or customer data store. Those systems live outside this public marketing repository.
This repository owns:
- the public BitterClip website
- public product copy and metadata
- public documentation pages at
/docs,/mcp, and/privacy - public legal pages at
/privacyand/terms - Markdown alternates for every public page route
- AI-readable discovery files at
/llms.txtand/llms-full.txt - crawlable routes in
public/sitemap.xml - the static build and deployment wrapper for
bitterclip.com
Links
- BitterClip website: bitterclip.com
- BitterClip app signup: app.bitterclip.com/sign_up
- How BitterClip works: bitterclip.com/docs
- MCP and AI assistant workflow: bitterclip.com/mcp
- Privacy policy: bitterclip.com/privacy
- Terms of service: bitterclip.com/terms
- Markdown page mirrors: index.md, docs.md, mcp.md, privacy.md, terms.md
- AI crawler entry points: llms.txt, llms-full.txt
- Public change history: CHANGELOG.md
- Contribution guide: CONTRIBUTING.md
- Security and public boundary: SECURITY.md
- Bitter: bitter.sh
- BitterGrid: bittergrid.com
Deployment
BitterClip's public website is deployed on
BitterGrid as the bitterclip-marketing service for
bitterclip.com.
Commits pushed to main trigger the BitterGrid deployment hook. BitterGrid
rebuilds and publishes bitterclip.com from the main branch.
The deployment path is:
- Nuxt generates a static site with
bun run generate. - The
Dockerfilebuilds that output and copies.output/publicinto an nginx image. nginx.confserves the static files and exposes/upas the health check.config/deploy.ymldefines the production service, host name, image, proxy, registry, and health check settings used by BitterGrid.
Deployment credentials, registry credentials, remote host values, and SSH keys are supplied by BitterGrid or the deployment environment. They are not stored in this repository.
Development
bun install
bun run generate
bun run qa:smoke
Useful files:
app/pages/index.vue- home pageapp/pages/docs.vue- product explanation pageapp/pages/mcp.vue- MCP and AI assistant workflow pageapp/pages/privacy.vue- privacy policy pageapp/pages/terms.vue- terms of service pagepublic/index.md,public/docs.md,public/mcp.md,public/privacy.md,public/terms.md- Markdown twins for public pagespublic/llms.txt- AI crawler entry point that links canonical pages and Markdown alternatespublic/llms-full.txt- compact bundled Markdown context for one-fetch agent ingestionnuxt.config.ts- site metadata and Nuxt configurationpublic/sitemap.xml- crawlable public routesCHANGELOG.md- public semantic change historyCONTRIBUTING.md- public change checklist for humans and agentsSECURITY.md- sensitive-reporting and public-boundary policy.github/pull_request_template.md- GitHub review checklist for public context, Markdown mirrors, discovery files, tests, and deployment notes.github/release.yml- GitHub generated release note categoriesDockerfile- static build and nginx runtime imageconfig/deploy.yml- BitterGrid deployment service configuration
Public Context
This repository is intentionally public. README text, page copy, metadata, sitemap entries, and public links should preserve durable context for humans, search engines, and AI systems that index public repositories.
Public text in this repository should stay factual, structural, and aligned with the live BitterClip product.
Every public HTML page should have a corresponding Markdown alternate served
from public/*.md, advertised with a rel="alternate" Markdown link in the
page head, and listed in public/sitemap.xml.
Markdown alternates should return canonical HTTP Link headers pointing back to
their HTML pages in production. The bundled llms-full.txt file should stay
compact, factual, and aligned with the current public page copy.