SDK plan (proposed)
Review the proposed TypeScript, Rust and Go client design and implementation gates.
A design plan, not a released SDK
The proposed first targets are TypeScript, Rust and Go. The Q4 commitment is to a design, examples and development plan; a released SDK remains a longer-term H1 2027 target. No installable package is announced here.
Read the SDK design and implementation plan for language API sketches, milestones, precision and pagination policies, and planned conformance cases.
For requests you can use today, open the GraphQL runner or the asset guide. The plan builds on existing account balances, asset holders and typed transaction workflows.
Preserve the API’s data boundaries
Planned clients must retain exact identifiers and raw amounts, source precision labels, coverage and freshness metadata, and route-scoped cursors. They must surface errors and schema incompatibilities instead of treating them as empty results.
The WebSocket publication feed remains origin-restricted first-party freshness hints. It does not provide a general wallet/asset subscription API or replayable event stream.
Future multihop pathfinding requires fresh venue state, asset identities, fees and precision, with distinct native Stellar and Soroban validation. The plan specifies these requirements; it does not claim an executable pathfinder.