Skip to content

Comments

docs: information architecture restructure#25

Merged
johnx25bd merged 18 commits intomainfrom
docs/information-architecture
Feb 24, 2026
Merged

docs: information architecture restructure#25
johnx25bd merged 18 commits intomainfrom
docs/information-architecture

Conversation

@johnx25bd
Copy link
Member

Summary

Restructures docs from blockchain-first framing ("geospatial policy engine for Ethereum") to wider "verifiable spatial computation" framing. Blockchain is one integration path alongside agents, APIs, and compliance systems.

  • Rewritten entry pages: introduction, quickstart, how-it-works lead with curl/API, not Solidity
  • 6 concept pages: geocomputation, geographic-features, result-format, privacy-model + edits to verifiable-computation, location-proofs
  • 6 guides: local-development, calling-the-api, blockchain-integration, agent-integration, building-plugins, verifying-location-proofs
  • 4 use cases: delivery-verification, parametric-insurance, geofence-compliance, onchain-attestation
  • 4 trust model pages: architecture, what-is-verified, what-you-are-trusting, security
  • API reference: compute/* and verify/* subdirectories
  • New: sdk/types, resources/research, resources/changelog

Known remaining work

  • ~9 pre-existing pages have old terminology (faq, playground, schemas, sdk/*, api-reference/overview)
  • Original replaced files preserved but not in navigation
  • Highest hallucination risk: agent-integration, building-plugins, verifying-location-proofs, privacy-model

Test plan

  • Review on Mintlify preview
  • Verify navigation links resolve
  • Check code examples against actual API/SDK
  • Flag fabricated claims or incorrect patterns

Adds RESTRUCTURE.md (rationale, audience analysis, migration plan)
and mint.proposed.json (new Mintlify navigation config).

Key changes: lead with verifiable computation not blockchain,
add trust model section, add agent integration as peer path,
relocate Ethereum-specific concepts into guides. Terminology
aligned with concept-map.md from docs-plan repo.
Replaces mint.json with proposed navigation structure:
- Getting Started, Core Concepts, Guides, Use Cases sections
- API Reference split into Compute, Verify, Records groups
- New Trust Model section
- Resources section with research and changelog
- Introduction: lead with verifiable spatial computation, not Ethereum
- Quickstart: curl-first approach, 4 clear steps, onchain is optional
- How-it-works: rename Policy Engine to Compute Engine, update diagrams

Reframes the entry point for a wider audience beyond blockchain
developers. Blockchain integration is positioned as one path among
several (agents, backends, compliance).
New pages:
- geocomputation.mdx (from geospatial-operations, terminology updated)
- geographic-features.mdx (from location-attestations, provenance section)
- result-format.mdx (from policy-attestations, universal framing)
- privacy-model.mdx (new: TEE privacy guarantees and limitations)

Edited:
- verifiable-computation.mdx (trust model content extracted)
- location-proofs.mdx (terminology: geocomputation, signed result)

Original files preserved for reference.
New guides:
- local-development: Docker, env setup, --env-file startup
- calling-the-api: request/response format, input types, errors
- blockchain-integration: EAS resolvers, delegated attestations
- agent-integration: Python/TS examples, decision-making patterns
- building-plugins: PoL plugin interface, Wi-Fi example
- verifying-location-proofs: claims, stamps, credibility scores

blockchain-integration merges content from eas-resolvers,
policy-attestations onchain sections, and location-attestations
storage options.
- delivery-verification: escrow + proximity resolver
- parametric-insurance: crop insurance with distance threshold
- geofence-compliance: drone corridor monitoring
- onchain-attestation: NFT minting, token distribution patterns

Each page uses story framing with concrete code examples.
delivery-verification relocated from guides/ with terminology updates.
Extracted from verifiable-computation.mdx and security.mdx:
- architecture: TEE execution model, container design, stateless model
- what-is-verified: signature coverage, input references, reproducibility
- what-you-are-trusting: honest trust assumptions, GPS spoofing caveat
- security: signer address, replay attacks, best practices, key rotation

Separates "how it works" from "what you should trust" for clearer
reasoning about the system's guarantees.
- api-reference/compute/*: relocate 6 operation pages to subdirectory
- api-reference/verify/*: split verify.mdx into stamp, proof, plugins
- sdk/types.mdx: TypeScript type definitions reference
- resources/research.mdx: links to location proofs paper, research agenda
- resources/changelog.mdx: v0.1.0 research preview entry
@johnx25bd
Copy link
Member Author

Code review

Found 8 issues:

  1. api-reference/overview.mdx card links point to old flat paths instead of new compute/ subpaths. Six cards link to /api-reference/distance, /api-reference/contains, etc. but mint.json now routes these to api-reference/compute/distance, api-reference/compute/contains, etc. Users clicking these cards land on orphaned pages with no sidebar context.

<CardGroup cols={2}>
<Card title="POST /distance" href="/api-reference/distance">
Distance between two geometries
</Card>
<Card title="POST /contains" href="/api-reference/contains">
Is geometry B inside geometry A?
</Card>
<Card title="POST /within" href="/api-reference/within">
Is point within radius of target?
</Card>
<Card title="POST /intersects" href="/api-reference/intersects">
Do geometries overlap?
</Card>
<Card title="POST /area" href="/api-reference/area">
Area of a polygon
</Card>
<Card title="POST /length" href="/api-reference/length">
Length of a line

  1. sdk/verify.mdx links to orphaned /api-reference/verify. The new nav splits this into api-reference/verify/stamp, api-reference/verify/proof, and api-reference/verify/plugins. The monolithic page is no longer in mint.json.

<Card title="API Reference: Verify" icon="code" href="/api-reference/verify">
View the HTTP API documentation

  1. how-it-works.mdx links to /concepts/eas-resolvers, which is not in the new navigation. RESTRUCTURE.md says this content moves to guides/blockchain-integration.

<Note>
Resolver contracts must be deployed and registered by the developer. Astral does not provide pre-built resolver contracts — you create them to implement your specific business logic. See [EAS Resolvers](/concepts/eas-resolvers) for patterns and examples.
</Note>

  1. use-cases/onchain-attestation.mdx links to /guides/geofenced-token, which is not in the new navigation. The file exists on disk but is orphaned from mint.json.

**Geofenced transfers** — restrict token transfers to users within a region (see the [geofenced token guide](/guides/geofenced-token) for the full implementation).

  1. API URL inconsistency across pages. The API reference documents /compute/v0/distance but the introduction, quickstart, and guide pages use /compute/distance (missing the v0 segment). Developers following the quickstart will hit a different URL than what the API reference documents.

```
POST /compute/v0/distance
```
vs
```bash
# How far is it from point A to point B?
curl -X POST https://api.astral.global/compute/distance \
-H "Content-Type: application/json" \
-d '{

  1. PR Add location proof plugin documentation #24 review feedback not addressed in api-reference/overview.mdx. Three changes were requested: (a) add Verify API to table/cards and reorder to Verify -> Compute -> Records, (b) change Records API base URL from /api/v0 to /records/v0, (c) change "across chains" to "across systems" in Records card. None were applied. (See PR #24 comments)

| API | Purpose | Base URL |
|-----|---------|----------|
| **Compute API** | Verifiable geospatial operations | `/compute/v0` |
| **Records API** | Query location attestations | `/api/v0` |
<CardGroup cols={2}>
<Card title="Compute API" icon="calculator" href="#compute-api">
Distance, containment, proximity checks with signed attestations
</Card>
<Card title="Records API" icon="database" href="/api-reference/records/overview">
Query existing location attestations across chains
</Card>
</CardGroup>
---

  1. concepts/result-format.mdx "Next" card links backward. The card points to /concepts/location-proofs, but in mint.json ordering, result-format comes after location-proofs. The "Next" link should point to /concepts/privacy-model.

<Card title="Next: Location proofs" icon="map-pin" href="/concepts/location-proofs">
Learn about verifiable location claims

  1. resources/schemas.mdx links to orphaned /resources/security. The security page was moved to trust-model/security but the "Next: Security" card in resources/schemas.mdx still points to the old path, which is not in mint.json.

<Card title="Next: Security" icon="shield" href="/resources/security">
Security considerations

🤖 Generated with Claude Code

- If this code review was useful, please react with 👍. Otherwise, react with 👎.

Rewrite introduction to lead with location proofs first,
geocomputation second. Show the full stamp pipeline (collect,
create, sign, compose) in both intro and quickstart examples.

Reorder Core Concepts in mint.json to follow progressive
conceptual build: geographic features → location proofs →
geocomputation → result format → verifiable computation →
privacy model.

Title case all page titles across concepts, guides, use cases,
and trust model sections.
Bring in PRs #27 and #28 (SDK + plugins overhaul, API reference
rewrite) as authoritative technical references
Previous version focused exclusively on the TEE/geocomputation side
and ignored location proofs. New version walks through the complete
pipeline: evidence collection → location proof composition →
verification → geocomputation → signed results.

Each step explains why it exists, not just what it does. Architecture
diagram shows both the location proofs path and geocomputation path
as composable but independent capabilities.
Three pages forming a progressive introduction to location data:
- location-data: parent page covering verifiability spectrum and data sources
- geojson: raw unsigned geometry, when to use it, coordinate conventions
- location-records: signed EAS attestations, LP v0.2 schema, storage options

Content migrated from geographic-features.mdx and location-attestations.mdx,
reorganized to separate concerns and build concepts progressively.
Six pages building the location proof concept progressively:
- location-proofs: parent page with certainty spectrum (L0-L5), key
  principles, and overview of the sub-pages
- pol-systems: categories of proof-of-location systems with honest
  assessment of current state
- location-stamps: atomic evidence units, independence from claims
- location-claims: assertions with required spatial uncertainty
- location-proof-structure: composition, single vs multi-stamp,
  plugin abstraction layer
- location-proof-evaluation: three-phase evaluation, credibility
  vector, confidence vs probability

Rewrites location-proofs.mdx from interface-heavy implementation doc
to conceptual explanation. No TypeScript interfaces or code snippets.
Four pages covering the geocomputation layer:
- geocomputation: parent page covering computational challenges,
  scope (verify + compute + future AI/ML), and verifiability
  approaches (TEE, ZK, smart contracts)
- astral-location-services: the hosted TEE service, verifiability
  and privacy properties, EigenCompute stack, signing key
- verify: the verification endpoint pipeline and its relationship
  to evaluation concepts
- compute: geospatial operations, accepted inputs, PostGIS/GEOS,
  precision/determinism, available and planned operations

Rewrites geocomputation.mdx from implementation-heavy to conceptual.
New pages separate the service (astral-location-services) from its
two capabilities (verify, compute).
Two standalone concept pages completing the Core Concepts section:
- signed-results: output formats, offchain and onchain usage,
  delegated attestation pattern, not every result needs blockchain
- privacy: TEE privacy guarantees, information leakage from results,
  spatial/temporal uncertainty as privacy tools, ZK research direction,
  honest TEE limitations

Content from result-format.mdx, policy-attestations.mdx, and
privacy-model.mdx reorganized into conceptual explanations without
code snippets or interface definitions.
Replace flat list of 6 concept pages with nested groups that tell a
progressive story when scanning the sidebar:

Location Data → Location Proofs → Geocomputation → Signed Results → Privacy

Each group has sub-pages that build on each other. Readers can scan
the sidebar and understand Astral's conceptual model without clicking
into any page.
Content migrated to the new Core Concepts structure:
- geographic-features → location-data, geojson, location-records
- location-attestations → location-records
- geospatial-operations → compute
- policy-attestations → signed-results
- eas-resolvers → removed (resolver patterns in guides)
- verifiable-computation → geocomputation, astral-location-services
- result-format → signed-results
- privacy-model → privacy

Also updates quickstart.mdx link to point to new location-data page.
Page-by-page review with hand-crafted edits throughout:

- Remove nonexistent confidence score from evaluation page,
  replace with accurate credibility vector structure from SDK
- Fix verify endpoint output (returns VerifiedLocationProof,
  not just credibility vector)
- Fix privacy page: v0 results include input plaintext,
  added warning and opened astral-location-services#57
- Fix GeoJSON page: Astral v0 accepts bare Geometry only,
  opened astral-sdk#67 for full spec parity
- Rewrite location proofs overview with certainty spectrum,
  design principles from research paper, neutral framework
- Generalize all pages: remove specific plugin references
  from conceptual content
- Render inline comments as callouts throughout
- Capitalize section headers consistently
- Add nav cards and see-also links to all pages
@johnx25bd johnx25bd merged commit e284d5a into main Feb 24, 2026
1 check passed
@johnx25bd johnx25bd deleted the docs/information-architecture branch February 24, 2026 03:29
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant