Ramoira: Open-source brand schema for claude, codex etc
Ramoira's free open brand schema drafts a valid 3.2.0 brand.schema.json with npx ramoira init and publishes Identity, Narrative, and Voice after review.
By Dillip Chowdary โข Oct 27, 2026 โข Source: ramoira.com
The news template wants two untitled intros and five two-paragraph sections. I'll lock the facts to the source, then measure each paragraph so the piece stays inside 800โ1,000 words.Ramoira is offering an open brand schema so every producer can check the same page, under the line "Your brand's meaning, written down and checkable," according to ramoira.com's report. Drafting is free and publishing is free. npx ramoira init installs ramoira/DRAFTING.md, the coding agent reads the project and asks what only the owner knows, and brand.schema.json is saved as a valid 3.2.0 schema. ramoira push sends that version for review before it is made public.
This piece is for solo builders, developers, brand founders, and agencies whose posts, sites, and ads disagree. It covers the 5-layer candidate schema, the local CLI, and the free public summary, including which layers a remote agent can fetch.
Open-source brand schema for claude, codex etc: what actually changed
Ramoira frames the same problem from four different angles. A solo builder says, "My AI content sounds different every time I post," because every session starts from scratch and early recognition requires consistency. A developer says, "I rebuild the same brand context for every project," because the brief never travels with the code and every handoff is a re-brief. A brand founder says, "My website says one thing. My ads say another," because the brand lives in the founder's head and every vendor reinterprets it slightly.
An agency says, "Our review cycles eat our production time," because violations happen when the spec is not machine-readable and every piece is a judgment call. The replacement is a 5-layer candidate schema, drafted by the owner's coding agent after npx ramoira init or in the browser at /draft, from the project and the owner's answers. brand.schema.json can be validated and checked locally against the spec. On ramoira push the owner affirms the rules and makes it public, as a candidate, not ratified.
Open-source brand schema for claude, codex etc: how it works

The open-source ramoira/cli drafts brand.schema.json locally and connects to Ramoira only when a version is pushed for review. npx ramoira init sets up drafting for the coding agent. ramoira check checks a draft against the owner's rules. ramoira validate validates it against the spec. ramoira push sends a version to ramoira.com for review. ramoira pull brings down the reviewed version. ramoira status checks publication status. No account is required to draft. Publishing requires a free account and includes everything in Draft.
Advertisement
Tech Pulse Daily
Get tomorrow's pulse first
Join engineers who read Tech Pulse before stand-up. Free, weekday mornings.
Publishing gives the schema a stable URL, and the brief travels with the work. Public rules plus the Identity, Narrative, and Voice layers are served on a cacheable endpoint, and the full schema stays private. The summary is at ramoira.com/brands/[slug]/. GET /brands/[slug]/schema.summary.json returns that cached slice so any agent can fetch it. GET /brands/[slug]/status returns publication state as facts only, never a score. The header is Cache-Control: public, max-age=300, stale-while-revalidate=3600. Agents, pipelines, and collaborators need no hosted file and no new re-brief.
Open-source brand schema for claude, codex etc: why it matters now
The CLI and the schema spec are entirely open source, so brand validation can sit in GitHub repositories, CI pipelines, and Cursor rules with no vendor lock-in. Every brand.schema.json on GitHub, every Cursor rules reference, and every CLAUDE.md belongs to the owner. The format works with Ramoira or without Ramoira. CLI source can be read after npx ramoira init, which Ramoira contrasts with a black box. The schema is connective tissue between a brand and the AI agents in use, and any tool, pipeline, or agent can support the open spec natively.
Open-source brand schema for claude, codex etc: who is affected
Draft is free with no account and works with Claude Code, Cursor, Windsurf, Replit, Lovable, and v0. Drafting is done by the owner's coding agent after npx ramoira init, or in the browser at /draft, as a 5-layer candidate schema from the project and the answers, checked locally against the spec. Published is free with an account. Each ramoira push is reviewed: the owner affirms the rules, then makes the schema public for remote agents and pipelines. The page labels that step sign in to publish.
Those tiers map to four roles. Solo builders need posts that stop sounding different every session, because early recognition requires consistency. Developers need brand context to travel with the code so a handoff is not a re-brief. Brand founders need the website and the ads to agree, rather than each vendor reinterpreting a brand that lives in someone's head. Agencies need the review cycles to stop eating production time once the spec is machine-readable and a piece is not only a judgment call.
Open-source brand schema for claude, codex etc: what to watch
The public brand stays a candidate, not a ratified brand. The owner affirms rules on ramoira push before the summary is public, and ramoira pull brings the reviewed version back down. ramoira status checks publication status, and GET /brands/[slug]/status is machine-readable and returns facts only, never a score. GET /brands/[slug]/schema.summary.json returns public rules plus Identity, Narrative, and Voice, while the full schema stays private. Cache-Control is public, max-age=300, stale-while-revalidate=3600, so a consumer can be served from cache inside that window.
The file is also meant to work with no hosted account. ramoira/cli runs locally until a push, and the source can be inspected. ramoira/brand-schema-spec supplies SPEC.md, fictional-brand examples, the reference validator, and per-layer docs for internal tools. ramoira/docs is written for agents and developers and names Cursor, Claude Code, Lovable, and v0, with integrations/cursor.md, integrations/claude-code.md, and integrations/v0.md. Windsurf and Replit are drafting clients alongside those tools. brand.schema.json, any Cursor rules reference, and the CLAUDE.md belong to the owner even if nobody pushes.
Developer Action Items
- โ Verify the claim on the official Claude / Codex page (or HN Claude/Codex/Fable), not from this recap alone.
- โ Name the surface that moved โ API, policy, model, hardware, or commercial terms โ before you Slack the thread.
- โ Assign one owner a day to read the primary material and decide: this-sprint, this-quarter, or noise.
- โ Do not change production on day-one coverage. Watch the vendor changelog and one independent write-up first.
Open-source brand schema for claude, codex etc FAQ
What is Ramoira's open brand schema?
It is a free, checkable brand.schema.json that records what a brand means. npx ramoira init installs ramoira/DRAFTING.md, a coding agent drafts a 5-layer candidate from the project and the owner's answers, and the file validates as schema 3.2.0.
Which tools can draft a Ramoira schema, and is an account required?
Drafting works with Claude Code, Cursor, Windsurf, Replit, Lovable, and v0, through the owner's coding agent after npx ramoira init or in the browser at /draft. No account is required to draft. A free account is required to publish.
What does ramoira push make public, and what stays private?
After the owner affirms the rules, Ramoira serves public rules plus the Identity, Narrative, and Voice layers at ramoira.com/brands/[slug]/ and at schema.summary.json. The full schema stays private, and the public brand is shown as a candidate, not ratified.
Which commands does the Ramoira CLI include?
npx ramoira init sets up drafting, ramoira check tests a draft against the owner's rules, and ramoira validate tests it against the spec. ramoira push sends a version for review, ramoira pull brings that reviewed version down, and ramoira status checks publication status. The CLI runs locally and connects to Ramoira only on push.
What does the Ramoira status endpoint return?
GET /brands/[slug]/status is machine-readable and returns publication state as facts only, never a score. Responses use Cache-Control: public, max-age=300, stale-while-revalidate=3600.
Sources
Author
Dillip Chowdary
Writes Tech Bytes coverage of AI, engineering, and the tools that actually ship. Editor of Tech Pulse Daily.
Related on Tech Bytes
Advertisement