MooseStack EOL Alternative
MooseStack is end of life. Here is the migration path.
MooseStack’s maintainers say the project is no longer actively maintained, and its GitHub repository is archived. hypequery replaces the typed ClickHouse query, semantic, API, React, and MCP layer while letting dedicated tools own schema, streaming, and workflows.
MooseStack status
End of life
hypequery model
Open-source TypeScript
Migration style
Incremental
An archived framework is not a new production dependency
MooseStack can keep running while you migrate, but new features and long-term application contracts need an actively maintained home.
Moose covered more than one job
Typed queries, DDL, Redpanda ingestion, Temporal workflows, APIs, and agent tooling should not be moved in one risky cutover.
The query layer can move first
Read-only ClickHouse queries and semantic metrics are the cleanest boundary to migrate and verify while existing infrastructure keeps running.
The replacement boundary
Move product analytics without rebuilding the data plane
hypequery gives the application an actively maintained, type-safe ClickHouse layer. Keep or replace ingestion, workflows, and DDL on their own timelines.
- Generate TypeScript types from the live ClickHouse schema
- Port and compare read queries one endpoint at a time
- Define shared dimensions, measures, metrics, and tenant rules in code
- Serve validated APIs and OpenAPI from your existing application
- Reuse contracts in React hooks and governed MCP tools
Step 1
Start from live ClickHouse
Physical schema stays with the migration system you choose. hypequery introspects the result and protects application queries with generated types.
Step 2
Migrate one query, prove parity, repeat
The old and new clients can coexist. Compare generated SQL and results before moving traffic, then promote shared calculations into datasets and metrics.
Use @hypequery/clickhouse for typed reads, @hypequery/datasets for shared meaning and tenant scope, and @hypequery/serve when the contract needs HTTP.
Assign Moose OLAP DDL, Redpanda, and Temporal responsibilities to explicit tools before removing the old runtime.
Typed read
A query inside your existing app
The result is inferred from the live schema, while filters and aggregations remain ClickHouse-native.
Migration map
What moves to hypequery—and what does not
This is a responsibility map, not a comparison between two active projects.
| MooseStack at EOL | hypequery | |
|---|---|---|
| Project status | End of life; repository archived | Actively maintained open-source packages |
| Typed queries | Existing Moose query code | @hypequery/clickhouse |
| Semantic metrics | Moose data models and APIs | @hypequery/datasets |
| HTTP and frontend | Moose query endpoints | @hypequery/serve and @hypequery/react |
| Agent access | Moose development harness | Governed @hypequery/mcp tools |
| DDL, streaming, workflows | Moose OLAP, Redpanda, and Temporal modules | Keep or choose dedicated tools |
Do not remove Moose until traffic, DDL, ingestion, jobs, and deployment dependencies have all been inventoried.
Where teams usually get stuck
Questions teams ask
MooseStack end of life
The official GitHub README states that MooseStack has reached end of life and is no longer actively maintained.
MooseStack migration guide
The migration guide separates physical schema, streaming, workflows, typed queries, APIs, React, and agents into safe stages.
Typed ClickHouse replacement
hypequery covers the application-facing TypeScript layer with generated schema types and ClickHouse-native queries.
Governed AI-agent analytics
Expose approved datasets and metrics through MCP instead of handing agents raw ClickHouse credentials.
Further reading
Go deeper where it actually helps
MooseStack migration guide
The staged technical and organisational cutover.
Open guide
hypequery vs MooseStack after EOL
The current status and responsibility comparison.
Open guide
Current hypequery capabilities
The shipped builder, semantic, Serve, React, and MCP surface.
Open guide
Quick start
Test one real ClickHouse table before planning the cutover.
Open guide
Next step
Inventory Moose responsibilities, then port one read query
Prove the typed ClickHouse path against production-shaped data before committing to the rest of the migration.