Developer Guides

Step-by-step developer guides for building with and on the PointSav platform. Each guide addresses a specific task — follow it start to finish, then refer back to the related architecture articles when you need the underlying theory.

For the concepts behind each guide, start in Platform architecture overview or Patterns. For platform architecture overview, see Totebox orchestration as the development environment.

Getting started

The foundation: authenticate your device, install the toolchain, and open your first session.

• Pair a new device — register a device with the pairing server and get it approved onto the network (now part of Machine Authorization)
• Install the development toolchain — set up Rust and the staging-tier commit helper on a workspace VM
• Open your first Totebox session — navigate to an archive, read your inbox, and start contributing
• Explore the console — tour the three-zone TUI, the status bar, and the F-key slots

Working in the console

Use the platform's terminal interface and its built-in Cartridges.

• Navigate the console TUI — the real screen layout and status bar fields
• Use the F-key model — what F3, F9, and F12 actually do
• Read the command ledger — page entries and fetch a signed checkpoint over service-fs's real HTTP API
• Run your first SLM query — the real path to a first inference request

Records & storage

Work with the WORM audit ledger and entity data.

• Read and write Totebox archives — session start reading protocol, commit flow, draft staging, cross-archive mailbox
• Verify a WORM ledger entry — verify against a fetched checkpoint using only curl and SHA-256
• Query the DataGraph — the real query_datagraph/get_entity_context tools
• Export structured data — three real export paths: DataGraph, wiki Markdown, ledger entries

Multi-entity scale

Manage multiple tenants, users, and fleet nodes.

• Configure a tenant namespace — the real config-driven provisioning, since no registration API exists
• Scale user access — grant role-scoped tokens as a team grows; there's no promote/revoke
• Add a node to a running fleet — enroll a second node without interrupting existing fleet operations

Integration & data

Connect external data pipelines and build location-intelligence applications.

• Build a co-location map — load a PMTiles archive directly; no REST API or API key exists
• Connect to the OSM data pipeline — the real ingest-osm.py script and taxonomy.py registration
• Federate archives via content mounts — mount a second repository's content into a running instance
• Use declarative knowledge mounts — the real mount schema and its real, unmitigated slug-collision risk

Self-hosting

Deploy the platform on your own infrastructure.

• Self-host a deployment — provision a Totebox Archive on a supported compute node
• Configure the Doorman gateway — set tier upstream addresses, circuit-breaker thresholds, and verify the health endpoint
• Deploy a knowledge instance — build and start app-mediakit-knowledge pointed at a local content repository
• Run local SLM inference — start the SLM service, verify Doorman Tier B, and submit inference requests from the console or API

See also

• Architecture — cross-cutting platform architecture
• Patterns — named design patterns used across the platform
• Totebox session — what a Totebox session is and what it can do
• Machine-based authorization — how machine-based authorization works
• Machine Authorization — device pairing, capability tokens, fleet enrollment, and binary-download authentication
• Self-Hosting — deploying platform components on your own infrastructure