# SemanticFed — product assets > 3 published assets for SemanticFed (3 files incl. ad sizes): documents, services, mockups. Source: https://burdenoff.com/products/semanticfed/assets ## SemanticFed — Product catalogue 2026 - Product: SemanticFed - Category: Documents - Format: pdf, pdf, 20 pages - File: https://assets.burdenoff.com/products/semanticfed/docs/semanticfed-brochure.pdf - Notes: https://assets.burdenoff.com/products/semanticfed/docs/semanticfed-brochure.md - View: https://burdenoff.com/products?assets=semanticfed&asset=semanticfed-brochure - Tags: brochure, product-catalogue, data-federation, semantic-layer, federated-query, data-governance, mcp, ai-agents, pricing 20-page landscape product catalogue for SemanticFed, a pre-launch data federation platform with a governed semantic layer, covering capabilities, AI agent surface, roles, security, plans and competitive positioning. ### About A 20-page, 16:9 landscape product catalogue (PDF title "SemanticFed — Product Catalogue") for SemanticFed, described as data federation with a governed semantic layer: queries run in place across databases, warehouses, lakes and business applications, against entities, dimensions and metrics defined once in a versioned model. The cover uses a dark navy design with teal accents and an illustration of data stores linked to a central hub; interior pages are mostly light-background card layouts. It is aimed at data platform leads, analytics engineers, data engineers, governance leads, analysts and AI/platform engineers. The document states the product is pre-launch (launching 2026, waitlist open) and labels capabilities as shipped, in progress or planned, including explicit notes on gaps such as run-time policy enforcement and agent confirmation gating. It is a product of Burdenoff Consultancy Services Pvt. Ltd. ### Contents 1. **Cover** — "Ask one question of all your data, answered where that data already lives." Tags: federated query, semantic layer, cross-source policy, agent-ready, query in place. Pre-launch notice; semanticfed.com. 2. **01 The short version** — Sources registered once; a planner pushes filtering and aggregation to each system and joins small results; semantic layer above. Promise: "Define it once. Query it where it lives." Figures: 1 definition per metric, 0 copies made, 40 agent tools, 2026 launch year. 3. **02 Who it is for** — Data platform leads, analytics engineers/BI leads, data engineers owning sources, governance/compliance leads, business analysts, AI/platform engineers; security architects mentioned. 4. **03 What breaks today** — Four problems with matching answers: data too scattered to query together, dashboards disagreeing, governance fragmenting per system, no provenance for numbers. 5. **04 What you get** — Nine governed objects: connectors, data sources (five-state lifecycle), connection health, semantic models (draft/published/deprecated/archived), entities and dimensions, metrics, relationships, saved queries and runs, access policies. 6. **05 Put to work** — Illustrative scenarios: analytics engineer joining live, warehouse and lake data; BI lead defining a measure once; compliance officer answering who queried what. 7. **05 Put to work, continued** — Data engineer replacing a copy pipeline with a virtual view; security architect keeping regulated data in-region; data architect viewing an entity/relationship graph for impact analysis. 8. **06 Built AI-native** — Agents bind to the published model; unknown fields fail before any source is contacted; agents get the caller's permissions; every answer opens into its plan; one permission-checked API behind agent surface, CLI and SDKs. 9. **07 What the AI can do, and what it cannot yet** — Shipped: 40 agent tools, two guided recipes, validation, stored plans. In progress: confirmation before irreversible agent actions; row/column policy at run time. Roadmap: natural-language to governed queries; model drafting from a source. States no model provider is wired in and the in-app AI page is a placeholder. 10. **08 What keeps it in bounds** — Nine guardrails: single permission edge, no agent-only data route, workspace isolation on writes, human publishing, human-only policy activation, propose-by-default loops, per-workspace off switch, source metadata treated as untrusted, nothing AI-only. 11. **09 Who does what** — Six seeded roles: Federation Owner, Governance Lead, Semantic Modeler, Source Engineer, Query Analyst, Auditor (across eleven resource types). 12. **10 Under the hood** — Metadata plane (shipped) and runtime plane (built; rough cost estimates, in-process execution, relational and REST drivers). Push-down execution, typed relationships, opt-in caching, runs as provenance. 13. **11 Standards, integrations and deployment** — GraphQL, Model Context Protocol, OIDC/OAuth 2.0 device code, SQL, Parquet; SQL wire endpoint on roadmap. Plugs into relational DBs, warehouses, object-storage lakes, REST/GraphQL apps, notebooks/pipelines, BI tools (once SQL endpoint lands). Deployment: managed multi-tenant, dedicated/self-hosted, hosting in India (Mumbai). 14. **12 Security and trust** — Single-edge authorisation, credentials by reference, results never stored, sparse roles, immutable policy rules, metadata-only sensitivity review. States no third-party certifications; SOC 2 and ISO 27001 on roadmap. 15. **13 One product, many front doors** — Shipped: MCP agent surface, command line, governed API. Planned: web console, Node and Python SDKs, editor/browser/desktop extensions. Docs at docs.semanticfed.com. 16. **14 Plans** — Free $0 (3 sources, 1 model), Starter $99/mo, Growth $399/mo (up to 250 sources), Business $999/mo (up to 1,000 sources), Enterprise ~$9 per source billed annually. Prices described as indicative; four regional bands including rupee pricing planned. 17. **15 Why teams choose it** — Federation plus semantic layer in one product, typed versioned model, stored plans, agent surface, flat fees, status labels on every capability. 18. **16 Honest answers to hard questions** — Pre-launch risk (recommends evaluation, not migration), policies stored but not enforced at run time, federated query performance caveats, exit/lock-in. 19. **17 Where it sits against the alternatives** — Comparisons with enterprise data virtualisation, distributed SQL engines, standalone semantic layers, warehouse/lakehouse platforms, hand-built pipelines and AI copilots on dashboards, including where those alternatives are stronger. 20. **Start today** — Suggests a short evaluation against two real systems; links: semanticfed.com/waitlist, semanticfed.com, semanticfed.com/ai, docs.semanticfed.com; company footer. ## SemanticFed — Services brochure - Product: SemanticFed - Category: Services - Format: pdf, pdf, 17 pages - File: https://assets.burdenoff.com/products/semanticfed/services/semanticfed-services-brochure.pdf - Notes: https://assets.burdenoff.com/products/semanticfed/services/semanticfed-services-brochure.md - View: https://burdenoff.com/products?assets=semanticfed&asset=semanticfed-services-brochure - Tags: services, consulting, data-federation, semantic-layer, implementation, migration, governance, managed-operations, training Services catalogue for SemanticFed: eight consulting, implementation, migration, governance, AI, managed-operations and training services across two tracks, plus how engagements run. ### About A 17-page landscape (16:9) services catalogue for SemanticFed, styled to match the product catalogue. It covers the consulting and managed federation services listed on semanticfed.com/services, in two tracks: "Improve your SemanticFed" (help adopting the platform) and "Powered by SemanticFed" (done-for-you outcomes delivered using it). It is for data-platform, analytics and governance leads deciding whether to bring in expert help with a SemanticFed rollout. No pricing is published, so the brochure points readers to the contact page. ### Contents 1. Cover: SemanticFed Services 2. 01 Expert hands on SemanticFed: overview and the two service tracks 3. 02 What we do: all eight services at a glance 4. 03 Track: Improve your SemanticFed (divider) 5. Federation Consulting & Architecture 6. Migration from ETL & Legacy Pipelines 7. Governance, Security & Compliance 8. AI & Semantic-Layer Enablement 9. Training & Team Enablement 10. 04 Track: Powered by SemanticFed (divider) 11. Implementation & Semantic Modelling 12. Integrations & Connector Setup 13. Managed Federation Operations 14. 05 Two ways to work with us: set it up with you, or managed federation operations 15. 06 How we engage: Discover, Plan, Build, Operate 16. 07 Enablement & support: handover, support tiers and SLAs, pricing (contact us) 17. Let's scope what you need: contact page, waitlist and services URLs ## SemanticFed — UI mockups - Product: SemanticFed - Category: Mockups - Format: pdf, pdf, 7 pages - File: https://assets.burdenoff.com/products/semanticfed/mockups/semanticfed-mockups.pdf - Notes: https://assets.burdenoff.com/products/semanticfed/mockups/semanticfed-mockups.md - View: https://burdenoff.com/products?assets=semanticfed&asset=semanticfed-mockups - Tags: mockups, ui, dark-theme, mcp, agents, semantic-model, query-plan, data-sources, access-policies, performance Seven dark-theme UI mockup screens of the SemanticFed web console: agent sessions, AI-proposed model, query run plan, data source health, tuning proposals, access policies and agent governance. ### About A 7-page PDF of full-page UI mockup images (no extractable text; each page is an image, 1200x670pt) showing the SemanticFed web console in a dark theme with a left navigation sidebar (Dashboard, Knowledge Graph, Ontology, Query Console, Agents, Topology, Registry, Policies, Monitoring, Configuration, Visualization, Audit Logs, Protocol, Collaboration, Analytics, Performance, Tech Stack, Deployment). All screens use a sample workspace called "Acme Data Platform". Several screens carry amber notices that state current gaps (policies not enforced at execution, no confirmation on destructive agent tools, in-app assistant not wired). The web console is planned, so these are design mockups rather than screenshots of a shipped UI. ### Contents 1. **Agent Sessions** — MCP clients connected to the workspace. Stat cards: 3 active sessions, 1,284 tool calls (24h), 40 tools available, 12 blocked by permissions. Recent tool calls table (models_publish, entities_create, queries_run, data_sources_transition, entities_delete blocked pending confirmation) from claude-code, cursor and a CI script. Session detail panel (client claude-code, transport stdio, 40 tools granted) and a warning "Destructive confirmation not enforced". "Generate Access Token" button. 2. **Proposed Model (Draft)** — Ontology / Revenue Core / Authoring Assist: a model generated from data source warehouse-prod (Postgres) with 3 entities (orders_fact, customers_dim, regions_dim), 7 dimensions, metrics (net_revenue SUM, order_count COUNT, avg_order_value AVG) and 2 MANY_TO_ONE relationships, with checkboxes, Discard / Accept Selected buttons, review notes (42 columns introspected, 16 objects proposed) and a note that publishing stays manual. 3. **Query run detail** — Run 8f41c2 "Active Revenue by Region — Revenue Core v4", status SUCCEEDED: 1,842 rows, 1.94s, 2 sources touched, estimated cost 2.10s. Execution plan with two steps (orders_fact with pushed predicates, projection and SUM aggregation; regions_dim), estimated vs actual rows, join order regions_dim to orders_fact. Plan summary panel and warning "Access policy not evaluated". 4. **Data Sources (Registry)** — Health and connectivity lifecycle: 6 connected, 1 degraded, 1 disconnected, 2,304 probes (24h). Table of sources (warehouse-prod, orders-primary, partner-rest-api, billing-replica, events-archive) with connector type (POSTGRES, REST_API, S3_PARQUET), status, last probe, latency and push-down. Expanded probe history for partner-rest-api (unreachable x4, upstream timeout). Proposed lifecycle transition panel with a note "Roadmap — proposals are not generated automatically yet". 5. **Tuning Proposals (Performance)** — Suggestions from run history: 48 queries analysed, 5 proposals open, est. 6m 40s saved per day. Cards for cache TTL changes (Active Revenue by Region, Daily Orders Snapshot) and a push-down proposal (Customer Tier Breakdown) with Dismiss / Apply buttons; side note "Nothing is applied automatically". 6. **Access Policies** — Proposed tab (3) listing column DENY rules on email, phone_number and date_of_birth, all inactive, with an expanded proposed policy for customers_dim.email (scope COLUMN, principal role:analyst, model Customer Core v2) and Approve & Activate / Reject buttons. Guardrails panel (proposals created inactive, activation requires a human, agents never toggle policies) and warning "Enforcement is not live yet". 7. **Agent Governance** — 40 tools exposed, 15 read-only, 6 destructive. Tool permissions table (models_list, queries_run, entities_create, models_publish, entities_delete, models_archive, access_policies_set_active) with confirmation status. Enforcement path diagram: Agent -> Workspace Gateway (permission check) -> Federation Service (workspace scoping) -> Data Source, "identical for human and agent callers". Open gaps: no confirmation on destructive tools; in-app assistant not wired.