Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.auvy.ai/llms.txt

Use this file to discover all available pages before exploring further.

​
Resources API

The Resources API is the unified discovery layer for AUVY objects. It reads from materialized resource heads in Mongo and gives clients one way to browse neurons, receptors, reflexes, pathways, meetings, stakeholder maps, assets, artifacts, and skills. Product-specific routes under /v1/pathways, /v1/neurons, /v1/skills, and similar remain supported CRUD and execution façades; they read and write the same underlying resource model. Use Resources when you need unified browse (list + resource trees), /v1/search, or POST /v1/resources/.../patch; use vertical routes when their payloads and workflows are a better fit.
Resource routes are workspace-scoped. Most reads also require a brain context via X-Brain-Id, a brain_id body field, or SDK getBrainId.

​
Browse (cross-kind lists)

For discovery across kinds, use POST /v1/browse. Set full_heads: true to receive materialized resource head documents in heads. Call once per type (e.g. neurons, pathways, receptors) or parallelize browse roots and merge client-side. 
POST /v1/browse
FieldTypeDescription
typestringBrowse root — same whitelist as hybrid search facets (neurons, pathways, …).
limitnumberOptional page size (1–200).
full_headsbooleanWhen true, response includes full heads array for list mode.
const neurons = await auvy.resources.browse.list({
  type: 'neurons',
  limit: 200,
  full_heads: true,
})

​
Vault Tree

The Vault tree merges native document folders with indexed resource heads in one response. 
POST /v1/browse/resource-tree
FieldTypeDescription
parent_keystringFolder/resource parent to open. Omit for root.
parent_keysstring[]Batch form; mutually exclusive with parent_key.
brain_idstringBrain UUID if not supplied by header/client default.
limitnumberNode cap per parent.
cursorstringCursor for single-parent pagination.
const root = await auvy.resources.browse.resourceTree({ parent_key: '/' })

const batch = await auvy.resources.browse.resourceTreeBatch({
  parent_keys: ['/Research', '/Artifacts'],
})

​
Document Tree and Moves

Document-lane folder utilities live under /v1/resources: 
GET  /v1/resources/tree
POST /v1/resources/tree
POST /v1/resources/move
tree returns row lists or folder shapes for path-indexed heads on the core spine. tree_profile: "document" (and the native document_tree tool) list artifacts, vault-document assets, and meetings. tree_profile omitted or broad is a wider catalog slice for HTTP/clients (not the assistant native tree tool). move updates folder paths by resource id; use move_profile: "document" for the same surface as native document_move (broader resource profile remains HTTP-only for operator flows).

​
Generic Reads

Some registry-backed resource kinds expose generic SDK list, read-by-id, and history paths. Neurons: list/history/version reads stay under /v1/resources/neurons/...; live head by resource UUID is GET /v1/resources/neurons/:resource_id (unified spine; SDK resources.catalogGet('neurons', resource_id)). GET /v1/neurons/:resource_id (client.neurons.get) returns the same neuron document shape; pass ?version=N for a full neuron-shaped historical read. Reflexes/receptors/stakeholder_maps also expose GET /v1/resources/:catalog/:resource_id for one-shot reads where implemented. 
GET /v1/resources/neurons
GET /v1/resources/neurons/:resource_id
GET /v1/neurons/:resource_id
GET /v1/neurons/:resource_id?version=N
GET /v1/resources/neurons/:resource_id/history
GET /v1/resources/neurons/:resource_id/versions/:version
The same generic route family is currently exposed for reflexes and receptors (list + read + history) and stakeholder_maps (as documented on each path). Receptor list and UUID read behavior matches GET /v1/receptors (including is_active and core-scope redaction); GET …/:id returns { receptor }. The @auvy-synapse/client receptors.list, receptors.getById, and receptors.get (when the second segment is a UUID) call these catalog routes. Slug-based reads and all mutations still use /v1/receptors/....

​
Central Patch

POST /v1/resources/:kind/:resource_id/patch
Central patch is intentionally narrow. It currently applies to artifact and meeting patch envelopes.

​
Live Resource Events

GET /v1/resource-events/stream
Server-sent events carry mutation cursors for resource creates, updates, deletes, and moves. Pass after=<redis-stream-id> to resume after a disconnect. 
import { subscribeResourceMutationStream } from '@auvy-synapse/client'

const stream = subscribeResourceMutationStream(auvy, {
  onEvent(event) {
    console.log(event.action, event.resource)
  },
})
  • Assets and Artifacts — document-lane rows
  • Browse — list-only by resource type
  • Search — hybrid keyword + semantic search (POST/GET /v1/search; former /v1/grep callers migrate here)