SurrealDB vs. data platforms
SurrealDB is not competing with Databricks, Snowflake, or other data platforms. We complement them. Different layers, different problems.
Data platforms are analytical warehouses for batch processing. SurrealDB is an operational database for real-time agent infrastructure. Your data platform answers "what happened." SurrealDB answers "what should happen next."Together, they form a complete data architecture.
COMPLEMENTARY
Different problems, different layers
Data platforms (Databricks, Snowflake) are analytical warehouses for batch processing. SurrealDB is an operational database for real-time agent infrastructure.
Data platforms
Batch analytics, data lakes, ML training pipelines. Optimised for historical analysis and model training.
SurrealDB
Real-time operations, agent context, transactional workloads. Optimised for serving and inference.
Together
Data platforms feed insights. SurrealDB serves them in real-time to agents and applications.
HOW THEY WORK TOGETHER
The data flow
Your data platform enriches features with batch analytics - customer segments, risk scores, demand forecasts. SurrealDB serves those enriched features in real-time to agents via live queries. The data platform answers “what happened.” SurrealDB answers “what should happen next.”
Ingest and enrich
Databricks or Snowflake processes raw data into features, segments, and scores through batch pipelines.
Serve in real-time
Enriched features flow into SurrealDB where agents and applications query them with sub-millisecond latency.
Agent context
Agents read enriched context from SurrealDB, reason over it with Spectron memory, and write decisions back - all in one transaction.
Feedback loop
Agent outcomes and user interactions flow back to the data platform for the next training cycle. Each cycle tightens retrieval quality and agent accuracy.
CONTEXT LAYERS
Context above the database vs. context in the database
A new category of “context layer” products is emerging - semantic platforms that sit between data sources and AI agents. They solve a real problem, but they solve it as middleware, not at the database layer.
Honeydew
Semantic layer compiling business logic into governed SQL for BI and AI agents. Strong governance, but static definitions that require manual maintenance.
Atlan
Enterprise data graph with catalog, lineage, governance, and MCP server. Rich metadata, but bootstraps context from existing systems rather than owning the transactional substrate.
RelationalAI
Decision intelligence platform inside Snowflake with graph, rules-based, and predictive reasoning. Powerful analytics, but tied to an analytical platform.
Puller AI
Self-service data retrieval with enterprise-grade semantic context. Useful for business users, but adds another intermediary between agents and data.
Credible
Context engine encoding shared meaning into Malloy-based semantic models. Delivers governed definitions, but cannot transactionally unify context with the underlying data.
Jedify
Semantic Fusion model merging multi-source data with business context via MCP/A2A servers. Aggregates context, but as a middleware layer above the databases.
SurrealDB's position: the context layer must live in the database, not above it. Only at the database layer can context be kept consistent, governed, secured, and transactionally unified with the canonical knowledge that grounds it.
CHOOSING THE RIGHT LAYER
Different problems, different tools
GET STARTED
Add the operational layer
