docforge/faq-schema.json at main · bdeeps/docforge · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
{
  "@context": "https://schema.org",
  "@type": "FAQPage",
  "name": "DocForge FAQ",
  "description": "Frequently asked questions about DocForge document ingest for RAG pipelines and AI agents.",
  "url": "/docs/faq.md",
  "mainEntity": [
    {
      "@type": "Question",
      "name": "What is DocForge?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "DocForge is a document ingest platform for RAG pipelines and AI agents. It converts PDF, HTML, DOCX, and text into clean Markdown and deterministic chunked JSON with token estimates. Agents self-register via REST or MCP and ingest through POST /api/v1/ingest or the ingest_document MCP tool."
      }
    },
    {
      "@type": "Question",
      "name": "What document formats does DocForge support?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "DocForge supports PDF, HTML, HTM, DOCX, TXT, MD, and Markdown. PDF uses PyMuPDF text extraction; HTML and DOCX are converted to Markdown; plain text and Markdown are normalized."
      }
    },
    {
      "@type": "Question",
      "name": "How do AI agents self-register with DocForge?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Via REST: POST /api/v1/agents/register with JSON body {\"name\": \"your-agent-name\"}. Via MCP: connect docforge-mcp and call register_agent. Save the api_key from the response — it is shown once only."
      }
    },
    {
      "@type": "Question",
      "name": "What is the difference between headings and size chunking?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Headings chunking splits on Markdown H1–H6 with heading_path metadata — best for structured docs. Size chunking uses fixed token windows with overlap — best for unstructured text and logs. Both are deterministic."
      }
    },
    {
      "@type": "Question",
      "name": "Is DocForge chunking deterministic?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Yes. Identical input and chunk configuration always produce the same chunk boundaries, content, and id values. Chunk ids are SHA256-derived for idempotent vector store upserts."
      }
    },
    {
      "@type": "Question",
      "name": "How do I set up DocForge MCP in Cursor?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Add docforge-mcp to Cursor MCP settings with DOCFORGE_API_URL (e.g. http://127.0.0.1:8787) and DOCFORGE_API_KEY from self-registration. Use cursor_config from the register response. Restart MCP and verify with health_check."
      }
    },
    {
      "@type": "Question",
      "name": "DocForge server not running — how do I fix it?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Run ./scripts/dev.sh or docforge-api after pip install -e .[dev]. Verify with curl http://127.0.0.1:8787/api/v1/health. Ensure DOCFORGE_API_URL in MCP config matches the running server."
      }
    },
    {
      "@type": "Question",
      "name": "Duplicate agent name error — what do I do?",
      "acceptedAnswer": {
        "@type": "Answer",
        "text": "Registration returns HTTP 409 when the agent name is already taken. Reuse your saved api_key, choose a unique name, or list agents via GET /api/v1/agents. API keys cannot be retrieved after registration."
      }
    }
  ]
}