Integrations · Hub

The tools your team already uses.

The workflow tools where work gets done, the experimentation platforms where work gets validated, the paid media platforms where spend gets steered, the data and analytics platforms where the questions get answered, the content and SEO platforms where the published surface gets produced, and the competitive-intelligence platforms that back the audit suite.

RampStack architecture diagram. A central navy hub card shows the RampStack mark with the subtitle 'Stack-agnostic methodology'. Six category cards radiate out: Workflow with 6 integrations (Jira, Linear, Notion, Figma, GitHub), Experimentation with 11 integrations (Statsig, PostHog, Optimizely, Amplitude), SEO Intelligence with 3 integrations (Ahrefs, Semrush, Similarweb), Paid Media with 5 integrations (Google Ads, Meta Ads, LinkedIn, TikTok, Synter), Content and SEO with 5 integrations (Webflow, Contentful, Frase, Profound, AirOps), and Data and Analytics with 5 integrations (BigQuery, Snowflake, Mixpanel, dbt, Hex).

The integration-orchestrator skill is the layer above. It produces a phased delivery plan, a 5-status taxonomy (Todo / In Progress / Waiting / Blocked / Done), and platform-specific commands per phase. The integration pages below show how the orchestrator's output lands inside each platform's real primitives.

View the orchestrator skill

The orchestration shape

Phases, gates, lock points.

The orchestrator skill outputs a five-phase delivery plan. Each phase ends in a gate; each gate locks the upstream decisions before downstream work begins. The shape is platform-agnostic; the integration pages below show how each platform represents phases, gates, and lock points in its own primitives.

Phased orchestration: Discovery, Direction, Identity, Production, Launch with gates between each phaseDiscoveryPhase 1DirectionPhase 2IdentityPhase 3ProductionPhase 4LaunchPhase 5Brief approvalgateDirection approvalgateIdentity approvalgateQA verificationgateLaunch readinessgatePhasesphasegate (lock point)
Five phases, five gates, five lock points. Jira holds this as Epics with workflow validators; Linear as Projects with cycle boundaries; Notion as Status transitions; Figma as locked variable collections; GitHub as required status checks plus CODEOWNERS.

The shape

Framework, integration, workflow.

Workflow & Project Management

The platforms where the brief lives during build.

Each microsite is platform-specific: Jira reads bureaucratic, Linear reads modern-startup, Notion reads docs-first, Figma reads designer, GitHub reads engineering. The agile page is platform-agnostic and covers how the brief lives across the sprint cycle no matter which tracker the team uses.

Experimentation & Feature Flags

The platforms where shipped work gets measured.

Ten platforms with first-party MCP servers covering experimentation, feature flags, and product analytics. Plus Eppo, included for completeness despite lacking an MCP today. Each microsite documents the MCP server URL, auth method, hosting model, scoping notes where relevant, and example prompts the agent will run.

These pages pair with the forthcoming experiment-design, feature-flagging, experimentation-analytics, and experimentation-platform-orchestrator skills. Microsite content stands on its own; skill cross-links land when the skill pages ship.

Data & Analytics

The platforms where the questions get answered.

Five platforms covering the modern data stack: BigQuery and Snowflake for warehouse, dbt for transformation and the Semantic Layer, Mixpanel for product analytics, Hex for collaborative exploratory work. All five ship official first-party MCPs (released 2025).

These pages pair with the forthcoming product-analytics-setup, data-warehouse-experimentation, and feature-launch-playbook skills. Microsite content stands on its own; skill cross-link hyperlinks land when the skill pages ship.

Content & SEO

The platforms where the published surface gets produced.

Five platforms covering the modern content stack: Webflow and Contentful for structured CMS publishing, Frase for the read-write SEO + GEO content lifecycle, Profound for AI search visibility measurement, and AirOps as a managed workflow alternative. All five ship official first-party MCPs released in 2025.

These pages pair with the forthcoming content-strategy-for-ai-search and content-production skills. Notion already lives in the workflow category; its existing microsite covers creative-direction briefs specifically and is unchanged. Microsite content stands on its own; skill cross-link hyperlinks land when the skill pages ship.

SEO & Competitive Intelligence

The platforms where the competitive landscape gets read.

Three platforms covering overlapping but distinct data shapes. Ahrefs is the foundational backend for the catalog's SEO audit suite (backlinks, keywords, content explorer, site audit). Semrush is the keyword and SEO-PR sibling with stronger US data and Topic Research. Similarweb is the standard for competitive traffic estimation, audience demographics, and channel-mix breakdown.

Use them in combination. Each platform gives a different cut of the same competitive landscape; the three often disagree on individual data points, and the disagreement is itself a useful signal.

Common stack combinations

Most teams run two or three of these together.

Pure single-platform shops are rare. The orchestrator skill outputs commands for the combination, not the single platform; the most common shapes are below.

  • Design-led brand work

    Notion + Figma + GitHub

    Brief lives as a Notion database row. Direction tokens and identity components live in a Figma library locked at each gate. Production code ships from a GitHub repo with brief-aware PR templates. All three MCPs run; gh CLI handles GitHub state.

  • Engineering-heavy

    Jira + Figma + GitHub

    Brief lives as a Jira Epic. Designs in Figma. Code in GitHub. Atlassian remote MCP plus Figma Dev Mode MCP plus GitHub MCP cover read paths. acli for Confluence content; gh for GitHub state.

  • Modern startup

    Linear + Notion + GitHub

    Brief lives as a Linear Project with a Notion long-form companion page; tickets live as Linear Issues; engineering happens on GitHub. Three MCPs, gh CLI for GitHub. The common combination for agent-led teams.

  • Docs-first agency

    Notion + Figma

    Briefs in Notion; visual library in Figma. No engineering surface. Notion MCP runs full-coverage; Figma Dev Mode MCP for read-side checks against the locked direction tokens. Common for brand consultancies and pre-engineering work.

  • Enterprise scrum

    Jira + Confluence + GitHub

    Brief lives in Confluence; Epics and Stories in Jira; code in GitHub. Atlassian remote MCP covers both Jira and Confluence; acli for content-heavy Confluence edits; gh for GitHub state.

  • Solo or two-person

    Notion + GitHub

    Brief in a Notion database; everything else in a GitHub repo. Notion MCP handles brief CRUD; gh handles repo and PR work. The minimum-viable shape the orchestrator skill still produces a useful plan for.

What every integration page does

The brief is the boundary object.

A creative-direction brief is the artifact that has to survive the team's default toolchain. The risk in every integration is the same: the brief gets produced, lives somewhere unreadable to the rest of the work, and stops being consulted. Six weeks later, the team is shipping competent and incoherent output, and no one can point at where the drift started.

Every integration page solves the same shape of problem, with platform-specific mechanics. The brief becomes a primitive the platform indexes (Epic, Project, database row, library description, repository document). The four axes become structured properties on that primitive (custom fields, labels, Select properties, token name prefixes, story categories). Downstream artifacts get queryable backlinks to the brief. The team's existing review ceremonies (sprint review, cycle review, design review, code review) gain one new question: does this answer the brief?

The mapping is what makes the brief load-bearing instead of decorative. Pick the integration page closest to the team's primary platform and adapt the patterns. The concepts transfer; the syntax differs.

In progress

Coverage we are still working on.

Slack, ClickUp, Asana, and Confluence are scoped for follow-up. The patterns transfer cleanly because the framework is the boundary object, not the tooling, but the platform-specific shape (Slack canvas vs ClickUp custom fields vs Asana sections) deserves its own page. If a platform is missing here that the team relies on, the existing pages translate well enough to ship a first draft of the integration internally.

Per-skill platform pages (a separate page for the brand-voice skill in Jira, the brand-voice skill in Linear, the brand-voice skill in Notion, and so on) are deliberately out of scope for this surface. The matrix gets large fast, and most of the value lives at the framework-to-platform level documented here.

Continue reading.

Frequently asked questions.

What is an integration page in RampStack?
A pattern document that explains how RampStack skill output (a creative-direction brief, a brand-voice doc, a brand-identity system) lands inside the primitives a specific platform already exposes (Jira Epics, Linear Projects, Notion databases, Figma libraries, GitHub repos). The page maps RampStack output to platform primitives, ships templates the reader can copy, and names the failure modes that show up when teams try the integration without the patterns.
Which platforms have integration pages right now?
Thirty-five, split across six categories. Workflow & project management: agile sprints (platform-agnostic), Jira, Linear, Notion, Figma, GitHub. Experimentation & feature flags: Statsig, PostHog, GrowthBook, Optimizely, VWO FME, Split.io (Harness FME), Kameleoon, LaunchDarkly, Amplitude, Flagsmith, plus Eppo. Paid media platforms: Google Ads, Meta Ads, LinkedIn Ads, TikTok Ads, and Synter (aggregator covering 14 platforms in one connection). Data & analytics: BigQuery (Google managed MCP), Snowflake (Snowflake-managed MCP with Cortex), Mixpanel (official hosted MCP), dbt (official dbt-labs/dbt-mcp), and Hex (official MCP). Content & SEO: Webflow (Data API + Designer API MCP), Contentful (local + hosted beta MCP), Frase (read-write MCP across the SEO + GEO content lifecycle), Profound (AI search visibility and Agent Analytics), and AirOps (managed workflow alternative with AEO data and Brand Kits). SEO & competitive intelligence: Ahrefs (foundational backend for the SEO audit suite), Semrush (sibling with stronger US keyword data and SEO-PR features), Similarweb (the standard for competitive traffic estimation, audience demographics, and channel mix).
Do I need RampStack to use these patterns?
No. The skills are open source under MIT and the patterns work whether the brief was generated by Claude using the creative-direction skill or written by a human using the same framework. The integration pages assume the brief exists; they do not assume how it got produced. Teams that already write briefs in some other format can translate to the four-axis format and gain queryable structure without changing tools.
How do these integration pages compose with the framework pages?
The framework pages explain the four axes and the brief format. The skill pages explain how to produce the brief and the downstream artifacts. The integration pages explain where those artifacts land inside the team's existing tools. Read framework first if the four-axis vocabulary is new. Read skills next if the question is what gets produced. Read integrations when the question is how to make the existing toolchain stop dropping the brief.