SDK v1.0, free app creation, and performance upgrades
The OnDB TypeScript SDK is now publicly available on npm, app creation is free, and the platform is significantly faster.New features
-
SDK publicly available on npm — The TypeScript SDK is now published as
@ondb/sdkunder an MIT license. Install it withnpm install @ondb/sdkand start building. Learn more - Free app creation — Creating a new app no longer requires a payment. You get your app ID and API keys immediately upon creation. Learn more
- Machine Payment Protocol (MPP) — A new payment method using Tempo for micro-USDC payments. The platform automatically discovers MPP support and handles settlement alongside x402. Learn more
- Public queries — Materialized views and predefined queries can now be marked as public, allowing unauthenticated access. Useful for building public-facing API endpoints. Learn more
- MCP tool discovery — Predefined queries registered as AI tools are now discoverable via schema endpoints, making integration with AI agents easier. Learn more
Updates
- Performance improvements — Significant latency reductions across the platform, including in-memory caching for shard metadata, index configs, and collection records. Read-heavy workloads should see noticeably faster responses.
- Multi-pod reliability — The payment ledger is now backed by Redis for safe operation across multiple server instances, eliminating potential double-spend issues at scale.
-
SDK security patches — Updated transitive dependencies to address known vulnerabilities in
axios,minimatch,picomatch,brace-expansion, andjs-yaml.
OnDB SDK v4.0 and platform launch
Major release of the OnDB TypeScript SDK and backend platform, delivering the missing infrastructure between proprietary data and AI agents.New features
-
Model API — A typed, declarative interface for structured reads and writes. Use
db.model<T>(collection)to getfindMany,findFirst,findUnique,create,createMany,updateDocument,deleteDocument,count, andaggregatemethods with full TypeScript support. Learn more -
Server-side JOINs — Execute
joinOneandjoinManyoperations on the backend in a single request. Supports nested JOINs and parent-field references with$data.fieldname. Learn more -
SQL interface — Query and insert data using SQL syntax via
db.sql()anddb.sqlInsert(). Learn more - Predefined queries — Create named, parameterized queries that execute publicly without authentication. Ideal for building custom API endpoints. Learn more
-
Materialized views — Define views over collections using JSON queries or SQL. Supports
live(auto-refresh on write) andlazy(manual refresh) modes. Learn more - Collection sharding — Partition large collections across shards using discrete, time-range, or hash-distributed strategies. Shard-aware query routing prunes irrelevant shards automatically. Learn more
-
Agent Keys — Keys with
Paypermission can pay other apps inline using USDC via EIP-3009. Configure spend limits, expiration, and target app whitelists. Learn more - Multi-chain payments — Pay for reads and writes across Celestia (TIA), EVM chains (Base, Ethereum, Polygon, Avalanche via USDC), and Solana using the x402 protocol. Learn more
- MCP server — Built-in Model Context Protocol server exposes your predefined queries as discoverable AI tools, with HTTP and stdio transports. Learn more
- Data retention policies — Configure per-collection retention with the first 30 days free. Monitor costs across all collections. Learn more
- Blob storage — Upload and retrieve binary files on-chain with metadata and task tracking. Learn more
Updates
-
Rebrand to OnDB — All SDK classes renamed from
OnChainDB*toOnDB*(e.g.,OnDBClient,OnDBConfig,OnDBError). -
Currency-agnostic pricing —
X402Quote.totalCostTiarenamed tototalCostto reflect multi-currency support (USDC, TIA, ETH, SOL). - Flexible payment routing — The SDK now matches the payment network returned by your callback instead of defaulting to the first option. You can choose any payment option the backend offers.
- Batch operations — Store multiple records in a single request with retry strategies for partial failures. Learn more
- Price index — Collections can define price fields with per-KB, per-record, or field-value pricing models with configurable revenue splits. Learn more
- Query builder operators — Full set of field operators including string matching, IP classification, GeoIP, CIDR range, regex, and keyword search. Learn more
-
Server-side aggregations —
count,sum,avg,min,max,distinctBy, andgroupByexecute on the backend. Learn more