Skip to content

API.md

Latest commit

 

History

History
264 lines (200 loc) · 4.27 KB

API.md

File metadata and controls

264 lines (200 loc) · 4.27 KB
Raw

VeilText HTTP API Reference

All API responses are JSON. JSON request bodies should use:

Content-Type: application/json

Typical error response:

{
  "ok": false,
  "error": "message"
}

Pipeline Steps

/api/encrypt and /api/decrypt use the same pipeline array. Each step needs a type; key-based algorithms may include key.

[
  { "type": "base64" },
  { "type": "aes_256_gcm", "key": "secret" }
]

Supported type values:

base16, base32, base58, base64, base85
aes_256_gcm, chacha20_poly1305, xchacha20_poly1305, aes_256_cbc
sha256, sha512, blake3, md5
js_hex_escape, js_unicode_escape, js_binary_string, js_jjencode, js_aaencode
js_jsfuck, js_eval_wrap, js_constructor_wrap, js_base36_tostring
bf_text, bf_emoji

Hash steps are one-way and cannot be reversed by /api/decrypt.

POST /api/encrypt

Runs the pipeline in order.

{
  "text": "hello world",
  "pipeline": [
    { "type": "base64" },
    { "type": "aes_256_gcm", "key": "secret" }
  ]
}

Response:

{
  "ok": true,
  "ciphertext": "...",
  "pipeline_desc": "Base64 -> AES-256-GCM",
  "plaintext_hash": "..."
}

Symmetric encryption outputs are Base64-wrapped for text transport.

POST /api/decrypt

Runs the provided pipeline in reverse order.

{
  "text": "...",
  "pipeline": [
    { "type": "base64" },
    { "type": "aes_256_gcm", "key": "secret" }
  ],
  "expected_hash": "optional sha256 hex"
}

Response:

{
  "ok": true,
  "plaintext": "hello world",
  "verified": null,
  "hash_match": null
}

If expected_hash is omitted, verified and hash_match are null; otherwise they are booleans.

POST /api/decode-smart

Automatically detects and decodes layered text.

{
  "text": "SGVsbG8",
  "key": "",
  "max_depth": 8,
  "pipeline": []
}

Response:

{
  "ok": true,
  "plaintext": "Hello",
  "pipeline_desc": "AI decode: undo Base64",
  "attempts": 1,
  "still_encoded": false,
  "steps": ["Base64"]
}

max_depth defaults to 8 and is clamped to 1 through 12. pipeline is optional and is tried first when provided.

Puzzle

POST /api/puzzle/split

{
  "text": "secret message",
  "pieces": 3
}

Response:

{
  "ok": true,
  "pieces": ["...", "...", "..."]
}

POST /api/puzzle/merge

{
  "pieces": ["...", "...", "..."]
}

Response:

{
  "ok": true,
  "text": "secret message"
}

POST /api/generate

{
  "template": "sk-{date:MMDD}-{word:tech}-{random:hex:8}"
}

Response:

{
  "ok": true,
  "text": "sk-0501-kubernetes-a3f12c9b"
}

Template variables are documented in TEMPLATES.md.

History

These audit endpoints are admin-only. Configure an admin token with --admin-token <token> or VEILTEXT_ADMIN_TOKEN, then send it as X-Admin-Token: <token> or Authorization: Bearer <token>.

Method Path Purpose
GET /api/history List backend audit history records
DELETE /api/history Clear backend audit history
DELETE /api/history/:id Delete one backend audit record

Template Data

Method Path Purpose
GET /api/template-data Read word banks and template data
PUT /api/template-data Replace template data
DELETE /api/template-data Reset template data to defaults

Settings

GET /api/settings

Returns basic UI settings:

{
  "ok": true,
  "settings": {
    "theme": "auto",
    "language": "auto"
  }
}

PUT /api/settings

Updates in-memory AI settings:

{
  "ai_provider": "openai",
  "ai_endpoint": "https://api.openai.com/v1",
  "ai_key": "sk-...",
  "ai_model": "gpt-4o-mini"
}

AI

POST /api/ai/test

Tests AI connectivity with the provided provider, endpoint, key, and model.

POST /api/ai/chat

AI assistant endpoint. The model can call built-in tools; the server performs up to 4 tool rounds.

{
  "message": "base64 encode hello",
  "provider": "openai",
  "endpoint": "https://api.openai.com/v1",
  "key": "sk-...",
  "model": "gpt-4o-mini"
}

Response:

{
  "ok": true,
  "reply": "...",
  "actions": []
}

Available tools:

encrypt, decrypt, generate, puzzle_split, puzzle_merge, web_search, get_time, system_info