Your AI agents are only
as smart as your context.
Meridian is the Context Platform that makes them useful.
A unified context layer, an orchestration engine, and a growing suite of agent-callable products. Connect it to Claude, ChatGPT, or your own agent — and let the work happen.
mcp.meridiansuite.ai · OAuth 2.1 + PKCE · Listed on Claude Connectors, ChatGPT Apps, PulseMCP
The thesis
“AI is a commodity. Context is differentiation.”
The model your agent uses is interchangeable. The customer relationships, voice, history, consent, and signals it acts on are not. Meridian gives every product — and every external agent — a single, governed context layer to act from.
The map
One governed substrate. Every agent. Every product.
The platform
Three layers. One coherent stack.
Context is the moat. Orchestration is the muscle. Applications are what you sell. Every layer is agent-callable from day one.
Context Layer
The moat.
- Unified contact graph across email, LinkedIn, WhatsApp, phone
- Voice DNA + drift detection
- Consent state, retention, regulatory zone
- Signal bus + event stream — the truth your agents need
Orchestration Layer
The execution muscle.
- Pattern matcher (ORCH-1) + super-agent dispatcher
- Workflow runner with cross-product invoke_agent
- Skill marketplace with no-code Skill Builder
- Action Board — every agent step visible and approvable
Application Layer
The products customers buy.
- Outreach OS, Content OS — live in production
- Media, Governance, Web, Engage, Analytics — coming
- Third-party apps snap on via MCP
- Every surface is agent-callable, not just human-clickable
MCP-native
Every product, callable by your agent.
mcp.meridiansuite.ai — SSE transport, OAuth 2.1 + PKCE, scoped tenant tokens, per-tenant rate and budget gates, kill switches at every level.
data_os.*NamespaceFoundational. Bundled with every customer.
query_segmentsget_contact_contextappend_eventresolve_identitycheck_consentlist_signalsoutreach.*NamespaceFor Outreach OS customers.
find_contactsscore_leaddraft_messagesend_messageschedule_campaignget_reply_intelligencerun_skillcontent.*NamespaceFor Content OS customers.
create_briefgenerate_draftscore_voicerepurposepublishdetect_authenticityoptimize_aeomeridian.*NamespaceCross-product platform plumbing.
invoke_agentpublish_handofflist_skillsrun_workflowThe flow
You type in Claude: “Find 50 fintech SaaS founders who engaged with our whitepaper, draft personalised outreach in our brand voice, schedule across LinkedIn and email at their best send-time, and queue the follow-ups.”
data_os.query_segments → outreach.find_contacts → data_os.get_contact_context → content.score_voice → outreach.draft_message → outreach.schedule_campaign
Meridian's super-agent orchestrates internally. The Action Board shows you exactly what's running. Approve a sample. Ship the rest.
The Meridian Suite
Eight products. One context layer.
Data OS is bundled with every Platform purchase — never a separate SKU. Every other product is agent-callable via MCP from day one.
DataOS
The Context Layer
Unified contact graph, identity resolution, consent, voice DNA, signal bus. Bundled with every product. Exposed to your AI via the data_os.* MCP namespace.
- Identity resolver
- Consent center
- Event explorer
- Segment builder
- Context inspector
- data_os.* MCP
OutreachOS
Agent-callable multi-channel outreach
8 AI agents, waterfall enrichment, 7-stage pipeline, 8-channel orchestration. Drives campaigns from your Claude or ChatGPT agent via outreach.* MCP.
- 8 AI agents
- 18 BYOK providers
- 7-intent reply engine
- outreach.* MCP
- Approval gates
- BYOK by default
ContentOS
Brand voice + authenticity + AEO/GEO
Brand Voice Engine with drift detection, authenticity ledger, AEO/GEO discoverability. 20 first-party skills including Runway, Heygen, Veo 3, ElevenLabs.
- Brand Voice Engine
- Skill marketplace
- Authenticity ledger
- AEO/GEO module
- content.* MCP
- 20 first-party skills
MediaOS
Paid media + creative + attribution
Cross-channel planning, creative production, MMM, and attribution. Built for Meridian's own go-to-market first, then sold externally.
- Cross-channel planning
- Creative production
- MMM + attribution
- Budget optimizer
GovernanceOS
AI authenticity + audit + compliance
Provenance per artifact, lineage, consent, audit trail, EU AI Act compliance. Primitives live in core today; commercialised later.
- Provenance ledger
- EU AI Act
- Lineage + audit
- Consent + retention
Web ExperienceOS
Agent-readable web + AEO delivery
Agent-readable CMS, schema.org JSON-LD, llms.txt, structured content APIs. Build sites that agents can actually read and act on.
- Agent-readable CMS
- llms.txt builder
- Schema.org JSON-LD
- Structured content APIs
EngageOS
Agentic decisioning at every touchpoint
A decisioning engine, not lifecycle messaging. Channels (email, SMS, push) are inputs to the agent, not the headline.
- Decisioning engine
- Real-time signals
- Multi-channel inputs
- Policy guardrails
AnalyticsOS
Revenue intelligence + agent-readable reports
Cross-product attribution, agent-readable narratives, revenue intelligence built on the Data OS context graph.
- Cross-product attribution
- Agent-readable reports
- Revenue intelligence
- Custom narratives
Creator OS is deferred to v2 until at least five products above are revenue-generating.
Why Meridian
Six things that make this real.
Not slideware. The Action Board, the skill marketplace, the cross-product super-agent — all in production today.
Action Board
Every agent step shown live — context fetched, tool called, draft scored. Approve a sample, ship the rest. Transparency by default.
Skill marketplace
20 first-party skills out of the box, including 14 media providers (Runway, Heygen, Veo 3, ElevenLabs, Nano Banana). Build your own with the no-code Skill Builder.
Cross-product super-agent
“Find 50 fintech founders, draft outreach in our voice, schedule across LinkedIn and email at best send-time.” One sentence. Multiple namespaces. One coherent run.
BYOK + budget gates
Bring your own keys. Per-tenant $10/day defaults, kill switches, tool-loop hard caps (8 calls / 60s / 4 iterations). Safety by construction.
Authenticity by default
Every AI-generated artifact carries a provenance record. EU AI Act compliance, content lineage, audit trail — embedded in the core, not bolted on.
Country-level compliance
GDPR, CCPA, PDPA, LGPD, APPI, DPDP, PIPA. Multi-currency billing. Timezone-aware operations. English at v1; Vietnamese next.
DataOS
The Context Layer that makes every agent useful.
Unified contact graph, identity resolution, consent, voice DNA, signal bus. Always bundled with Platform — never a separate SKU. Exposed to your AI via the data_os.* namespace.
Consent Center
Per-contact consent state, retention, regulatory zone (GDPR/CCPA/PDPA/LGPD/APPI/DPDP/PIPA). One-click export and deletion.
Identity Resolver
Merged identity graphs across email, WhatsApp, LinkedIn, phone with manual merge/split and audit trail.
Event Explorer
Real-time stream per contact with filter, search, and pivot to context card.
Segment Builder
Drag-drop UI on top of the segment infrastructure every product depends on.
Context Inspector
“See what your AI sees.” Show exactly what bundle a customer's agent receives on a contact. The differentiator screen.
Voice DNA Engine
Brand voice extraction, scoring, and drift detection across every artifact, every channel, every agent.
Integrations
Bring your own keys. Cascade by cost.
Free scrapers run first. Paid providers cascade only when needed. Every integration is BYOK.
Data providers (waterfall)
Channels
Media + model providers
Pricing
Bundle what you need.
Platform base (Data OS + Auth + Integration Hub) is always included. Add the products you actually use.
Scale + 4
Talk to us
Data OS + four products. Skill marketplace publishing. Custom rate caps.
Get started →Enterprise
Talk to us
Everything. Dedicated MCP region. Governance OS. EU AI Act tooling.
Get started →BYOK pricing model. Per-tenant $10/day default budget gates. Usage-priced add-ons for premium media skills.
Give your AI agents a context layer worth using.
Connect Meridian to Claude or ChatGPT. Run the demo. Ship the work.