This guide covers the optional organisation service. Configuration, console and API paths refer to your Lolly Work instance. The packaged console keeps its release-matched copy of this guide.
Your team finishes a poster, and it has to reach the campaign archive exactly as approved. This page is for the admin who sets up those fixed organisation destinations in Lolly Work, and for the people and automations that send to them. Delivery sends one finished Lolly export to a destination your organisation owns: S3-compatible storage (such as AWS S3, MinIO, Ceph RGW or Garage), WebDAV (including Nextcloud and ownCloud), an HMAC-signed HTTPS receiver, Brandfolder or Cloudinary.
How delivery differs from a source
A catalog provider brings remote assets into Lolly, normally read-only. Delivery sends a generated output out. Reusing one credential for both would silently turn read authority into write authority, so the two keep separate contracts, registries and secrets.
- Personal targets stay personal. A person's own send targets stay on their device. Lolly Work cannot see, disable, redirect, credential or choose them, and they never appear in Lolly Work configuration, grants, history or credentials.
- Organisation credentials stay on the server. A shell receives only
{id, kind, label, formats, maxBytes, visibility, projectFiles?}, after group and permission filtering. - Side by side, never instead. The connected shell shows organisation destinations beside personal ones. An organisation S3 destination cannot replace a person's S3 connection or become their automatic export home, and a person's preference cannot withdraw an organisation destination.
- Nothing by default. No destination is configured until you add one, and no destination means no outbound delivery.
- Explicit sends. A manual send is a person's choice. Automation may publish only its own completed, unchangeable render output by reference; it cannot choose a personal file, another principal's job or an arbitrary storage key.
- Signed exports only. An export send needs a Content Credential on the exact bytes sent. DAM destinations may also accept a verified shared-project original with
projectFiles: true; local and federated catalog assets never gain that authority. - No lock-in. S3 is an open protocol adapter, not a privileged hosted service. You bring your own endpoint, bucket and key, and can leave without migrating a Lolly-specific format.
Pick an adapter
| Adapter | Credential | After writing, Lolly Work checks | Revoke |
|---|---|---|---|
s3 | <accessKeyId>:<secretAccessKey> | a signed HEAD confirms the size | no |
webdav | <username>:<app-password>, or bearer:<token> | authenticated HEAD and a bounded GET confirm size and SHA-256 | no |
https | a dedicated random HMAC secret | the receiver's returned digest, when it sends one | no; the protocol has no safe universal delete |
brandfolder | an API token used only for writing | reads the original back and compares SHA-256 | no; it never updates, replaces, moves or deletes an asset |
cloudinary | JSON { "apiKey", "apiSecret" } | reads the original back and compares SHA-256 | no; an id collision is never replaced |
S3, WebDAV and signed HTTPS refuse redirects, and Brandfolder refuses API and upload redirects. S3 and WebDAV receipts record transformation: "none" and the delivered digest. Signed HTTPS derives that status from the receiver's digest rather than assuming every publisher keeps bytes unchanged.
Configure a destination
Destinations live in the instance configuration under delivery. Every destination has an id, a kind, a label, a credentialRef (the environment variable that holds its secret), enabled, the groups that may see it, the formats it accepts and an optional maxBytes below the instance-wide delivery.maxBytes.
Use a write credential dedicated to one destination and restricted to its bucket, folder or section. Do not reuse a catalog-provider credential, the instance BlobStore credential or an operator's broad release key.
Configure a fixed S3 destination
{
"delivery": {
"maxBytes": 67108864,
"destinations": [
{
"id": "campaign-archive",
"kind": "s3",
"label": "Campaign archive",
"credentialRef": "LW_DESTINATION_CAMPAIGN_ARCHIVE",
"enabled": true,
"groups": ["marketing", "brand"],
"formats": ["svg", "png", "jpg", "pdf"],
"maxBytes": 33554432,
"options": {
"bucket": "approved-output",
"endpoint": "https://minio.example.org",
"region": "us-east-1",
"prefix": "campaigns"
}
}
]
}
}Set the referenced secret in the deployment environment:
LW_DESTINATION_CAMPAIGN_ARCHIVE='<accessKeyId>:<secretAccessKey>'publicBaseUrl is optional and means exactly what it says: when present, Lolly Work marks the destination public and returns the matching object URL in the receipt. An unguessable public URL is not access control.
Configure a fixed WebDAV destination
The collection, and any prefix subdirectory, must already exist. Delivery neither browses the account nor creates folders; it writes only beneath the fixed collection URL.
{
"id": "team-files",
"kind": "webdav",
"label": "Approved team files",
"credentialRef": "LW_DESTINATION_TEAM_FILES",
"enabled": true,
"groups": ["brand"],
"formats": ["svg", "png", "pdf"],
"options": {
"url": "https://cloud.example/remote.php/dav/files/team/outgoing",
"prefix": "approved"
}
}Set LW_DESTINATION_TEAM_FILES='<username>:<app-password>', or bearer:<token> for a token-fronted DAV server. For Nextcloud, use a revocable app password, not the account password. A predictable flat filename makes a retry overwrite the same resource. publicBaseUrl is optional and separate from the credentialed collection URL.
Configure a signed-HTTPS destination
This adapter connects an organisation-owned publishing service without tying Lolly Work to its vendor or API. The target is one exact HTTPS endpoint, not a webhook a person types in:
{
"id": "press-publisher",
"kind": "https",
"label": "Press publisher",
"credentialRef": "LW_DESTINATION_PRESS_PUBLISHER",
"enabled": true,
"groups": ["communications"],
"formats": ["png", "jpg", "pdf"],
"options": { "url": "https://publisher.example/lolly-delivery" }
}Set the referenced variable to a dedicated random HMAC secret. Lolly Work sends the raw export body with x-lolly-delivery-id, x-lolly-format, a base64url x-lolly-name, x-lolly-content-sha256, x-lolly-timestamp and x-lolly-signature: v1=<hex-hmac-sha256>. The signed canonical string is:
lolly-delivery-v1\n<timestamp>\n<delivery-id>\n<format>\n<base64url-name>\n<content-type>\n<size>\n<sha256>The receiver should reject stale timestamps and duplicate delivery ids. It may return no body, or JSON { "id": "…", "url": "https://…", "sha256": "…" } (at most 64 KiB). A matching digest records transformation: "none", a different digest records "provider-managed", and no digest records "unknown".
Configure Brandfolder writes
Use a separate write credential, even when Brandfolder is already a catalog source: the read-only source does not grant delivery permission. Configure one existing Brandfolder and file section per destination:
{
"id": "brandfolder-campaigns",
"kind": "brandfolder",
"label": "Brandfolder campaigns",
"credentialRef": "LW_DESTINATION_BRANDFOLDER",
"enabled": true,
"groups": ["brand"],
"formats": ["png", "jpg", "pdf"],
"projectFiles": true,
"options": {
"brandfolderId": "your-brandfolder-id",
"sectionId": "your-section-id",
"visibility": "private",
"uploadOrigins": ["https://your-qualified-upload-origin.example"],
"downloadOrigins": ["https://your-qualified-original-origin.example"]
}
}- The referenced variable holds the Brandfolder API token. Bearer credentials go only to
https://brandfolder.com/api/v4, never to storage. - Supply the exact HTTPS storage origins you observed and checked for your tenant; the example origins are placeholders. Reading the original back follows at most three redirects, checking the exact download origin at every hop, without API credentials.
visibilitydeclares the existing Brandfolder's access policy; it does not change that policy. Check the setting again when administrators change remote access.
The adapter follows Brandfolder Binary Upload: it requests a temporary upload ticket, uploads the original, then creates one asset in the chosen section. Brandfolder's temporary object URL is reachable by its importer and managed by the vendor. As of October 2026, Brandfolder's documentation says upload URLs expire after 24 hours and temporary objects after seven days. Lolly Work never publishes its own source blob or returns those temporary URLs to the shell.
A private progress record is saved before creation and again with the created asset ID, so a processing or verification failure retries that known asset without uploading or creating it again. If the create response is lost, Brandfolder documents no idempotency key, so the delivery stops with an uncertain-outcome message. An operator must reconcile the asset before another create is attempted; do not create a fresh delivery just to get past the message. The adapter does not assign collection membership.
Configure Cloudinary writes
This adapter stores authenticated raw originals in a product environment that uses dynamic folders. It does not generate image or video renditions or publish a public CDN URL.
{
"id": "cloudinary-originals",
"kind": "cloudinary",
"label": "Cloudinary originals",
"credentialRef": "LW_DESTINATION_CLOUDINARY",
"enabled": true,
"groups": ["brand"],
"formats": ["png", "jpg", "svg", "pdf", "mp4"],
"maxBytes": 67108864,
"projectFiles": true,
"options": {
"cloudName": "your-cloud",
"region": "eu",
"folderMode": "dynamic",
"assetFolder": "campaigns",
"publicIdPrefix": "lolly-deliveries"
}
}- The secret is JSON
{ "apiKey": "...", "apiSecret": "..." }, used only on the server. Regions areus(default),euandap. - The folder and the public-ID prefix are set separately:
asset_folderdoes not imply a public-ID path in dynamic-folder mode. Fixed-folder tenants are not supported. - Keep the delivery limit at or below 64 MiB; chunked uploads are outside this adapter.
- Uploads use an operation-specific public ID,
type=authenticatedandoverwrite=false. A retry first looks up that ID through the Admin API, and accepts existing bytes only when the destination, operation markers and original download digest all match. - Only the delivery ID and SHA-256 are added as reconciliation context. No source tags, approval, private notes, prompts or other organisation fields are copied. The adapter sends no upload preset, no eager or incoming transformation and no moderation, AI, add-on or URL-import parameters. Check tenant-wide processing settings separately. See the Upload API and folder modes.
Both DAM adapters are fixture-tested. Before enabling one, use a dedicated test section or folder to check credentials, tenant settings, readback, access, processing time and timeout recovery on your own tenant.
Send a shared original
With projectFiles: true on a DAM destination, the Assets view offers that destination when the selected asset is a shared-project original. The shell posts an empty body to /api/v1/destinations/:id/deliveries?projectId=...&fileId=...&checksum=...; no name or format override is accepted. The server reads the ready file, verifies every stored part and the whole digest, and works out the format from the bytes, with the stored filename as a fallback. The destination's format list and size limit apply, and no browser download and upload round trip is needed.
- The original stays in its project. Lolly Work stages an unchangeable private copy for retry, and keeps the source reference only on the server.
- The configured submit scan hook runs before the file leaves, with the same reject and error policy as catalog admission. No hook means no scan was performed.
- The source person's current project access, account status, destination grants and configuration are checked again before sending, after scanning, on retry and after any human approval. Service tokens cannot choose project originals.
- The destination's own access rules apply. Sending an original does not copy its project permissions or approval state.
What happens when you send
A raw body is accepted only when it carries Lolly's C2PA export assertion. Before anything leaves, Lolly Work records its SHA-256 and stages the exact bytes in the configured BlobStore.
- awaiting approvalonly when the destination has an approval chain
- queuedthe exact bytes are staged
- deliveringone attempt at a time, across replicas
- delivereda receipt with the delivered digest
- failed, then retryThe same staged bytes and, for S3, the same key.
- rejected or cancelledA reviewer rejects, or the sender withdraws. Final: retry cannot bypass the decision.
- Interrupted while deliveringHeld for an operator to reconcile, never taken over automatically.
- Destination changedRetry answers
409 DESTINATION_CHANGED; make a new delivery.
A delivery waits for approval when its destination has a chain, then is queued, delivered or failed. A retry sends the same staged bytes again; a reject or a withdrawal ends the delivery.
signed by Lollywork-delivery-lifecycle.htmlHTMLpixels, not shapes11 KBAI generatedgenerated by Claude Opus 5.5Check it yourselfGet the signed file- The record does not change. It binds the destination version, name, format, MIME type, size and SHA-256, in both the memory and PostgreSQL stores.
- One attempt at a time. Attempts are claimed atomically across replicas, so two retries cannot both run. A process interrupted while
deliveringstays held for an operator to reconcile; another worker does not take it over automatically. - Retries send the same bytes. For S3, a retry uses the staged bytes and the same key:
<prefix>/<delivery-id>/<sha256-prefix>-<safe-filename>.<format>. - A changed destination needs a new delivery. If the destination's endpoint, bucket, prefix, formats, exposure or other meaning changes, its version changes and a retry returns
409 DESTINATION_CHANGED. Credential rotation does not change the version. - History. Delivery history is scoped to the principal that sent the delivery. Stored records and audit events contain destination ids and output facts, never credentials. Creation is audited once as
delivery.created, and each attempt recordsdelivery.deliveredordelivery.failed. - Rate limits. Creating and retrying share the automation rate-limit bucket, because each may spend an organisation credential; discovery and receipt history do not.
- Removing a destination stops new sends and retries. It does not delete remote objects or erase receipts.
Delivering an automation output
POST /api/v1/jobs/:id/deliveries sends a completed render job owned by the same member or service principal, and POST /api/v1/renders/:id/deliveries does the same for a succeeded durable render. The JSON body is { "destinationId": "campaign-archive", "name": "Launch poster" }; an optional format must equal the recorded format and cannot override that format. Both routes run one code path, and a durable render also checks that its owner still holds the tool and export.server.
When the job completes, Lolly Work records a SHA-256 over the actual output bytes, independently of the storage provider (an S3 ETag is not a portable content digest). Delivery reads the stored result, checks it against that digest, verifies its Lolly Content Credential again, then points the delivery's private sourceRef at the existing blob. There is no download and upload round trip and no second stored copy. Jobs completed before digests were stored must be rendered again before this route can deliver them.
The receipt shows sourceJobId or sourceRenderId for correlation, never the private storage reference. The source job cannot be deleted while any delivery keeps its output (409 JOB_OUTPUT_IN_USE), even after a successful send, so retry and audit cannot be undermined by deleting their evidence. A durable render's output retention leaves alone any output a delivery points at, for the same reason.
Optional human approval
Set approvalChain on a destination to an existing chain id. Creation then returns 202 with state: "awaiting-approval" and an approvalId, and nothing has been sent. The approval subject is the delivery id, whose record already binds the destination version, name, format, size and digest, and a final approval sends that exact staged record.
- A rejection becomes the final delivery state
rejected, and a withdrawal becomescancelled. Neither can be relabelled as a transport failure or bypassed through the retry route. - A configured chain that does not exist fails closed, before the request body is accepted.
- Service tokens cannot create a human-review request, for raw bodies and job outputs alike. Future workflow automation must enter through its own explicit approval step.
How chains work is on Approvals.
Who may send
delivery.create is the right to send to a fixed organisation destination. Members carry it by default, because the operator has already chosen, credentialed and enabled each destination. The destination's groups and resource grants narrow it further:
lw grants add group:contractors delivery.create 'destination:campaign-archive' --effect deny- A matching deny removes that destination from the caller's
org-config, and is checked again when the delivery is written. The same deny-wins decision applies to discovery, creation and retry. - An explicit per-destination allow can extend a destination beyond its configured groups, for example to give a
viewerservice token exactly one delivery right without raising its role. - When a destination has an
approvalChain,delivery.createallows staging and requesting review, not the send itself; the approval chain and its separation-of-duties rules decide that. - For an automation output, the same action adds no general storage authority: the caller must own the job or durable render, it must be a finished render, its recorded format cannot change and its stored bytes pass integrity and C2PA checks again.
- Losing the permission stops new sends and retries, but does not erase the caller's own receipts.
delivery.createhas no authority over personal send targets.
Grants in general are on Roles and grants.
API example
curl -X POST \
-H 'Cookie: lw_session=…' \
-H 'Content-Type: image/png' \
-H 'Idempotency-Key: campaign-42-poster-v1' \
--data-binary @poster.png \
'https://work.example.com/api/v1/destinations/campaign-archive/deliveries?name=Launch%20poster&format=png'Discovery and history routes are in the API reference.
Check a destination against a live system
The S3 adapter has been checked against a live S3-compatible host. WebDAV and signed HTTPS have injected-wire tests, and opt-in checks in the Lolly Work repository that have not yet run against a real tenant or receiver, so neither is called live-verified.
The opt-in interoperability checks
Three gated tests in the Lolly Work repository check each adapter against a real endpoint. Each is skipped unless its switch is set to 1. It writes inert text under a random canary-<uuid> delivery id and removes that exact object afterwards, or fails and lists each id whose removal it could not confirm. None of them lists or browses the remote side or prints a credential.
| Test | Switch | Settings (each prefixed with the switch name) |
|---|---|---|
tests/s3-upcloud-canary.test.ts | LW_S3_CANARY | _ENDPOINT, _REGION, _BUCKET, _ACCESS_KEY, _SECRET_KEY |
tests/webdav-delivery-canary.test.ts | LW_WEBDAV_CANARY | _URL, _CREDENTIAL, optional _PREFIX |
tests/https-delivery-canary.test.ts | LW_HTTPS_CANARY | _URL, _SECRET, _CLEANUP_URL, optional _CLEANUP_CREDENTIAL |
The S3 test accepts any S3-compatible endpoint. Point the WebDAV test at an existing collection set aside for the test: it delivers twice to confirm that a retry overwrites the same resource, then reads the bytes back with its own request.
The signed-HTTPS protocol has no delete contract, so that test deletes through the receiver's own management API. LW_HTTPS_CANARY_CLEANUP_URL is an HTTPS URL with one {id} placeholder, filled with the id from each receipt, and a GET of the same URL must then answer 404 or 410. When a receipt is missing, unusable or has no id, the test uses the delivery id instead and counts the item as removed only when the DELETE succeeds. If the management API cannot resolve delivery ids, the test reports the cleanup as unconfirmed, and the operator must check the receiver by hand. When the receipt carries a digest, it must match the bytes sent. A second HTTPS test sends three requests the receiver must refuse with a 4xx status: a signature from another secret, a timestamp a day old and a body that does not match its signed digest.
Not built yet
- A versioned small-workflow document and runner above the job-output send.
- A replica-safe background claimant for queued or stalled deliveries.
- Retention for staged retry bytes, and for job outputs a delivery keeps.
- Destination setup in the console or CLI; destinations are configuration only.
- A passing run of the WebDAV and signed-HTTPS checks against a real tenant and receiver.
Any further DAM destination must declare how it writes (create, new version or replace) and its conflict policy before it writes anything. Transfer acceptance, processing, original readiness, review and public activation need separate observations, and an uncertain completion must be reconciled against the remote asset and its committed effects before another write. Approval of a source cannot stand in for approval of the staged final digest. These are requirements for a new destination, not capabilities of the current adapters.