drime-api
Knowledge about Drime Cloud API endpoints, authentication, file operations, workspaces, vault encryption, and S3 upload flow. Use when working with API integration, adding new endpoints, or debugging API issues.
📖 中文导读:Knowledge about Drime Cloud API endpoints, authentication, file operations, workspaces, vault encryption, and S3 upload flow. Use when working with API integration, adding new endpoints, or debugging API issues. 以下正文为该技能的原始 SKILL.md 内容(英文),上方安装命令可直接使用。
name: drime-api description: Knowledge about Drime Cloud API endpoints, authentication, file operations, workspaces, vault encryption, and S3 upload flow. Use when working with API integration, adding new endpoints, or debugging API issues.
Drime Cloud API Integration
This skill provides knowledge about the Drime Cloud REST API for the shell application.
API Base
- Production:
https://app.drime.cloud/api/v1 - Authentication:
Authorization: Bearer <token>
Core Concepts
IDs vs Hashes
- Numeric IDs: Used for mutations (move, delete, copy, rename)
- Base64 Hashes: Used for downloads and shareable links
- Hash calculation:
base64(id + "|")with trailing=stripped
Workspace ID
0= Default personal workspace- Always pass
workspaceIdquery param for scoped operations
Pagination
- Use
perPage=9999999999to fetch all in one request - Simplifies client logic, API handles large result sets
Key Endpoints
Authentication
GET /cli/loggedUser → User info
POST /auth/login → { email, password, token_name } → { user, access_token }
File Listing
GET /users/{id}/folders?workspaceId={id}&perPage=9999999999 → All folders (tree)
GET /drive/file-entries?parentIds[]={id}&workspaceId={id} → Children of folder
File Operations
POST /folders → Create folder { name, parentId?, workspaceId }
POST /file-entries/move → { entryIds[], destinationId?, workspaceId }
POST /file-entries/duplicate → Copy files
POST /file-entries/delete → { entryIds[], deleteForever?, emptyTrash? }
PUT /file-entries/{id} → Rename { name }
S3 Upload Flow
Uploads use a 4-step process to offload bandwidth to S3/R2.
Upload Constants
const (
ChunkSize = 60 * 1024 * 1024 // 60MB per part
MultipartThreshold = 65 * 1024 * 1024 // Use multipart above 65MB
BatchSize = 8 // Sign URLs in batches of 8
PartUploadRetries = 5 // Retry individual parts
)
Flow
Validate:
POST /uploads/validate- Checks for duplicates and quota
- Handle duplicates via policy:
ask,replace,rename,skip
Presign:
- Small (<65MB):
POST /s3/simple/presign→ Single PUT URL - Large (>=65MB):
POST /s3/multipart/create→uploadId+key- Then
POST /s3/multipart/batch-sign-part-urlsfor part URLs
- Then
- Small (<65MB):
Transfer: PUT binary to presigned URL(s)
- For multipart: collect ETags, complete via
POST /s3/multipart/complete - On failure:
POST /s3/multipart/abort
- For multipart: collect ETags, complete via
Finalize:
POST /s3/entries- Creates FileEntry in database
- Requires:
filename,size,clientName,relativePath
Hash Calculation
The Drime hash is computed from the file ID:
// Encode: ID → Hash
func CalculateDrimeHash(fileID int64) string {
hashStr := fmt.Sprintf("%d|", fileID)
encoded := base64.StdEncoding.EncodeToString([]byte(hashStr))
return strings.TrimRight(encoded, "=")
}
// Decode: Hash → ID
func DecodeDrimeHash(hash string) (int64, error) {
padding := (4 - len(hash)%4) % 4
hash += strings.Repeat("=", padding)
decoded, _ := base64.StdEncoding.DecodeString(hash)
idStr := strings.TrimSuffix(string(decoded), "|")
return strconv.ParseInt(idStr, 10, 64)
}
Vault Encryption (Client-Side)
Zero-knowledge encryption using AES-256-GCM:
- Key Derivation: PBKDF2-SHA256 (250k iterations) with user password + vault salt
- Encryption: AES-256-GCM with random 12-byte IV per file
- Auth Tag: 16 bytes appended to ciphertext
- Uploads: Encrypt before S3 upload, set
isEncrypted=1inPOST /s3/entries - Downloads:
GET /file-entries/download/{hash}?encrypted=true
Vault Endpoints
GET /vault → Metadata (salt, iv, check)
POST /vault → Initialize { password, password_confirmation }
GET /vault/file-entries → List vault entries (?parentHash={hash})
POST /vault/delete-entries → Permanent delete (no trash in vault)
Workspaces
GET /me/workspaces → List workspaces
GET /workspace/{id} → Details with members
POST /workspace → Create { name }
PUT /workspace/{id} → Update
DELETE /workspace/{id} → Delete
POST /workspace/{id}/invite → { email, roleId }
DELETE /workspace/{id}/member/{id} → Kick member
GET /workspace_roles → Available roles
Sharing
POST /file-entries/{id}/shareable-link → Create public link
PUT /file-entries/{id}/shareable-link → Update settings
DELETE /file-entries/{id}/shareable-link → Remove link
POST /file-entries/{id}/share → { emails[], roleId } (email invites)
Starring & Tracking
POST /file-entries/{id}/star → Star
DELETE /file-entries/{id}/star → Unstar
PUT /file-entries/{id}/add-tracking → Start tracking
DELETE /file-entries/{id}/tracking → Stop tracking
GET /track/infos/{id} → Get tracking stats
Error Handling
Error Response Codes
| Code | Meaning | Client Action |
|---|---|---|
401 Unauthorized |
Token expired or invalid | Prompt re-login, clear stored token |
403 Forbidden |
Permission denied | Check workspace membership, entry permissions |
404 Not Found |
Entry doesn't exist or no access | Verify path, check cache staleness |
429 Too Many Requests |
Rate limited | Read Retry-After header, back off |
500+ Server Error |
Backend issue | Retry with exponential backoff |
Client-Side Error Handling
// Check for token expiration
if errors.Is(err, api.ErrTokenExpired) {
return fmt.Errorf("session expired, please run: login")
}
// Retry transient failures
func withRetry(fn func() error) error {
for i := 0; i < 3; i++ {
if err := fn(); err != nil {
if isTransient(err) {
time.Sleep(time.Duration(i+1) * time.Second)
continue
}
return err
}
return nil
}
return fmt.Errorf("max retries exceeded")
}
Reference
See drime-openapi.yaml for complete API specification.