Zeer Chef
Zeer Chef turns Zcash data into application entities with TypeScript handlers, PostgreSQL storage and a GraphQL API.
These guides target Chef and CLI 0.7.0. The Prep client has its own version, 0.1.0. Your project's schema and installed SDK determine its API; the hosted Playground retains its separately deployed demo schema.
How Chef turns chain data into your application API.
Chef defines and stores your application model. Prep supplies the public block and transaction data; applications can also query Prep directly.
Key features
- TypeScript handlers — transform blocks, transactions and authorized receipts.
- Entity schemas — define the records stored in PostgreSQL.
- Relationships and lists — connect records and query bounded nested results.
- GraphQL — query committed entities and checkpoint metadata.
- Recovery — resume checkpoints and handle supported reorgs.
- Shielded & Memos — read authorized notes through a local testnet scanner.
Performance Benchmarks
Recorded 20 October 2026 over blocks 0–9,999 inclusive. Chef started with an empty project namespace and produced 10,000 BlockSummary entities, with a separate GraphQL process serving reads during indexing. Durations are in seconds (s).
| Chef input profile | First data observed in GraphQL | 10,000 entities indexed |
|---|---|---|
| Fixed source pages, 32-block commits | 1.201 s | 36.696 s |
| Adaptive source pages, 32-block commits | 1.591 s | 35.215 s |
Each profile has one recorded run. The timer includes CLI startup, Prep reads, handler execution, PostgreSQL commits and polling through completion. Both runs reached checkpoint 9,999 with identical entity results.
These are historical block-summary measurements. The later Chef 0.6.1 address-analytics study separately measures a 512-block application replay with concurrent GraphQL reads. The 0.6.0 acceptance record verifies the runtime's package installation, transaction breakdown and restart behavior; 0.6.1 retains that runtime. Chef 0.7.0 operational acceptance separately verifies the heartbeat, diagnosis and source-interruption workflow.
See the Chef benchmark methodology and recorded samples for timing boundaries, resource limits and evidence.
To tune your own project, start with Database / Queries, Latency at the Head or Benchmarking, depending on the delay you observe.
Use Cases
- Application backends — Turn chain data into your own entity model and give a frontend an authenticated GraphQL API.
- Activity dashboards — Use block handlers and transaction handlers to store counts and summaries for charts, reports or historical analysis.
- Specialized explorers — Model transaction inputs and outputs as related entities, with exact values and explicit unresolved history.
- Monitoring and notifications — Poll bounded entity ranges from a separate consumer and deliver notifications after applying your recovery policy. Chef does not provide subscriptions, webhooks or a durable change feed.
- Payment inboxes and relayers — Run the local shielded scanner to index readable notes, exact received amounts and memos. Your application matches receipts to orders and checks confirmations and canonicality before authorizing a payout.
Quick links
Quickstart · Quickstart with AI · Benchmarks · Playground · GitHub