Question 1

What is Metalogue?

Accepted Answer

Metalogue is the Universal Context Layer for Enterprise AI. It connects your organization's knowledge sources (Slack, Notion, GitHub, Jira, and 90+ more) to any AI agent, enabling federated semantic search without migrating your data.

Question 2

How is Metalogue different from a vector database?

Accepted Answer

Vector databases store embeddings. Metalogue federates across your existing infrastructure. You keep your data where it is—we provide the unified query layer. No migration, no vendor lock-in, and inherited permissions from source systems.

Question 3

What is the Ghost Graph?

Accepted Answer

Ghost Graph is Metalogue's ephemeral context assembly system. It builds temporary knowledge graphs for each query, connecting relevant context from multiple sources without persisting sensitive relationships. The graph dissolves after use.

Question 4

Do I need to migrate my data to use Metalogue?

Accepted Answer

No. Metalogue connects to your existing data sources where they live. We sync metadata and embeddings, but your documents stay in Slack, Notion, GitHub, or wherever you store them. Zero migration required.

Question 5

What AI agents work with Metalogue?

Accepted Answer

Metalogue exposes context via MCP (Model Context Protocol). It works with Claude, ChatGPT, Cursor, custom agents, and any system that can call our REST API or MCP endpoints.

Question 6

How quickly can I get started?

Accepted Answer

Most teams are up and running in under 15 minutes. Install our SDK, connect your first data source, and start querying. Our average time-to-first-query is 12 minutes.

Question 7

Is there a free tier?

Accepted Answer

Yes. Our Starter tier includes 10,000 queries per month, 3 data sources, and core features—free forever. Perfect for small teams and proof-of-concept projects.

Question 8

What programming languages do you support?

Accepted Answer

We offer official SDKs for TypeScript/JavaScript, Python, and Go. Our REST API works with any language that can make HTTP requests.

Question 9

What is Vec2Vec translation?

Accepted Answer

Vec2Vec is Metalogue's neural translation layer. It converts embeddings between different vector spaces (OpenAI ↔ Cohere ↔ local models), enabling cross-model search without re-embedding your entire corpus.

Question 10

How does federated search work?

Accepted Answer

Metalogue queries multiple data sources in parallel, applies permission filtering at the source level, and merges results using our semantic fusion algorithm. Sub-Dijkstra pathfinding optimizes context traversal.

Question 11

What embedding models do you support?

Accepted Answer

We support OpenAI (ada-002, text-embedding-3), Cohere (embed-v3), Anthropic, Google (gecko), Mistral, and custom/local models. Vec2Vec translation enables cross-model compatibility.

Question 12

How do source permissions work?

Accepted Answer

Permissions are inherited from source systems. If a user can't access a Slack channel or Notion page, they won't see those results in Metalogue. We verify permissions at query time, not sync time.

Question 13

What is the MCP endpoint?

Accepted Answer

MCP (Model Context Protocol) is an open standard for providing context to AI models. Metalogue exposes your organizational knowledge via MCP, enabling Claude, Cursor, and other AI tools to directly query your context.

Question 14

How do you handle real-time updates?

Accepted Answer

We use webhook listeners and polling (configurable) to detect changes in source systems. Updates typically reflect in search within 30 seconds for webhook-enabled sources, or 5 minutes for polling.

Question 15

Can I use my own vector database?

Accepted Answer

Yes—BYOV (Bring Your Own Vector). Professional and Enterprise tiers can connect their existing Pinecone, Weaviate, Qdrant, or Milvus. We query your infrastructure directly; zero data leaves your control. Starter/Free tiers use our Managed Intelligence tier with tenant-isolated enclaves.

Question 16

What is Semantic Unlearning?

Accepted Answer

Semantic Unlearning is our GDPR/CCPA compliance feature. It uses anti-vectors to mathematically remove information from your context space without re-indexing. Deletion requests complete in under 100ms.

Question 17

How do I integrate with my existing RAG pipeline?

Accepted Answer

Metalogue can serve as your retrieval layer, replacing or augmenting your existing vector store queries. Our SDK drops into LangChain, LlamaIndex, and custom pipelines with minimal code changes.

Question 18

How does pricing work?

Accepted Answer

Metalogue uses a three-part model: Base license (per node), usage-based queries (per 1K), and optional Shield add-ons for enterprise compliance. This aligns your costs with actual value delivered.

Question 19

What counts as a query?

Accepted Answer

A query is a single search request to the Metalogue API. Federated searches that span multiple sources still count as one query. Sync operations and admin API calls don't count toward your quota.

Question 20

Is there a volume discount?

Accepted Answer

Yes. Usage tiers provide automatic discounts: Standard ($0.02/query), High Volume ($0.015), Enterprise ($0.01), and negotiated rates for 10M+ monthly queries.

Question 21

What's included in the Enterprise tier?

Accepted Answer

Enterprise includes SSO/SAML, SCIM provisioning, dedicated support, SLA guarantees (99.9%), custom retention policies, and advanced audit logs. SOC2 Type II attestation coming Q2 2026.

Question 22

Can I try before I buy?

Accepted Answer

Absolutely. Our Starter tier is free forever with 10K queries/month. For Enterprise features, we offer a 14-day free trial with full access. Contact sales for custom POC arrangements.

Question 23

How do overages work?

Accepted Answer

We don't cut you off. If you exceed your plan's included queries, we bill overages at the next tier's rate. You'll receive alerts at 80% and 100% usage to avoid surprises.

Question 24

Is Metalogue SOC2 compliant?

Accepted Answer

SOC2 Type II certification is in progress and expected Q2 2026. We follow SOC2 controls today and can share our security posture documentation upon request. Contact us for our current security questionnaire.

Question 25

Do you support GDPR compliance?

Accepted Answer

Yes. We're fully GDPR compliant with data residency options (US, EU), DPA availability, and Semantic Unlearning for right-to-erasure requests. We act as a data processor under GDPR.

Question 26

Is Metalogue HIPAA eligible?

Accepted Answer

Yes. Our Enterprise tier supports HIPAA compliance with BAA availability, PHI handling procedures, and encrypted storage. Contact us to discuss healthcare deployments.

Question 27

How is data encrypted?

Accepted Answer

Managed Sovereignty: Your data stays in tenant-isolated vector enclaves. AES-256 at rest, TLS 1.3 in transit. Enterprise customers get Customer-Managed Encryption Keys (CMEK)—you control the keys, you can mathematically destroy the data. BYOV customers: we never store anything; we query your Pinecone/Weaviate directly.

Question 28

Where is my data stored?

Accepted Answer

By default, US-East (AWS). EU customers can request EU-West data residency. Enterprise customers can specify custom regions or bring their own infrastructure.

Question 29

Do you access my data?

Accepted Answer

Metalogue employees cannot access customer data without explicit permission. All access is logged, and we support customer-controlled audit logs for compliance reviews.

Question 30

How do you handle data retention?

Accepted Answer

Default retention is indefinite with soft-delete. Enterprise customers can configure custom retention policies (30/60/90/365 days). Semantic Unlearning enables immediate, verifiable deletion.

Question 31

What SSO providers do you support?

Accepted Answer

We support SAML 2.0 and OIDC. Native integrations with Okta, Azure AD (Entra ID), Google Workspace, OneLogin, and Ping Identity. Custom IdP integration available.

Question 32

Do you support SCIM provisioning?

Accepted Answer

Yes. SCIM 2.0 for automated user and group lifecycle management. Users are automatically provisioned, updated, and deprovisioned based on your identity provider changes.

Question 33

What is your SLA?

Accepted Answer

Enterprise tier includes 99.9% uptime SLA with service credits for violations. We publish real-time status at status.metalogue.xyz and historical uptime reports.

Question 34

How does enterprise onboarding work?

Accepted Answer

You get a dedicated Solutions Engineer for integration planning, a Customer Success Manager for ongoing support, and priority access to our engineering team for custom requirements.

Question 35

Can I get a custom contract?

Accepted Answer

Yes. Enterprise customers can negotiate custom terms including payment schedules, liability caps, indemnification, and SLA modifications. Contact sales@metalogue.xyz.

Question 36

Do you offer professional services?

Accepted Answer

Yes. Our Professional Services team offers integration consulting, custom connector development, security reviews, and architecture design. Rates vary by scope.

Question 37

What support channels are available?

Accepted Answer

Starter: Community Discord, docs. Pro: Email support (24hr SLA). Enterprise: Dedicated Slack channel, phone support, and named account manager.

Frequently Asked Questions

Quick Answers

Still have questions?