Integrations

Unstructured Compatibility

AILANG Parse exposes a drop-in compatible endpoint at POST /general/v0/general that accepts the same multipart file upload format as the Unstructured API. Existing Unstructured clients can point at AILANG Parse with just a URL change.

curl

# Same request format as Unstructured API
curl -X POST https://docparse.ailang.sunholo.com/general/v0/general \
  -F "files=@report.docx" \
  -H "Accept: application/json"

Python — before and after

# Before: Unstructured
from unstructured.partition.auto import partition
elements = partition(filename="report.docx")

# After: AILANG Parse (same API surface)
import requests
resp = requests.post(
    "https://docparse.ailang.sunholo.com/general/v0/general",
    files={"files": open("report.docx", "rb")}
)
elements = resp.json()

The response format follows the Unstructured element schema — each element has a type, text, and metadata field. For Office formats (DOCX, PPTX, XLSX), AILANG Parse extracts structural features that Unstructured misses: track changes, merged cells, text boxes, comments, and headers/footers.

With Docling

Docling excels at PDF layout analysis with its DocLayNet model. AILANG Parse excels at deterministic Office parsing. Use both — route by file type:

• DOCX, PPTX, XLSX, ODT, ODP, ODS, CSV, HTML, Markdown, EPUB → AILANG Parse (free, fast, deterministic)
• PDF, scanned images → Docling (layout analysis, table detection)

Python routing example

import requests
from pathlib import Path

AILANG_PARSE_URL = "https://docparse.ailang.sunholo.com"
OFFICE_EXTS = {".docx", ".pptx", ".xlsx", ".odt", ".odp", ".ods",
               ".csv", ".html", ".md", ".epub"}

def parse_document(filepath: str) -> dict:
    ext = Path(filepath).suffix.lower()

    if ext in OFFICE_EXTS:
        # AILANG Parse: deterministic, no per-page billing, full structural fidelity
        resp = requests.post(
            f"{AILANG_PARSE_URL}/api/v1/parse",
            files={"file": open(filepath, "rb")}
        )
        return resp.json()

    elif ext in {".pdf", ".png", ".jpg", ".jpeg", ".tiff"}:
        # Docling: layout-aware PDF and image parsing
        from docling.document_converter import DocumentConverter
        converter = DocumentConverter()
        result = converter.convert(filepath)
        return result.document.export_to_dict()

    else:
        raise ValueError(f"Unsupported format: {ext}")

This gives you structural Office parsing from AILANG Parse and layout-aware PDF parsing from Docling.

With LlamaParse

LlamaParse is a cloud LLM-powered parser — strong on PDFs, but every file costs API credits. AILANG Parse handles Office formats locally for free, saving your LlamaParse budget for PDFs that actually need LLM analysis.

• Office formats → AILANG Parse (free, local, deterministic, no per-page bill)
• PDFs → LlamaParse (cloud, LLM-powered, paid per page)

Python routing example

import requests
from pathlib import Path

AILANG_PARSE_URL = "https://docparse.ailang.sunholo.com"
OFFICE_EXTS = {".docx", ".pptx", ".xlsx", ".odt", ".odp", ".ods",
               ".csv", ".html", ".md", ".epub"}

def parse_document(filepath: str) -> dict:
    ext = Path(filepath).suffix.lower()

    if ext in OFFICE_EXTS:
        # AILANG Parse: free, local, deterministic
        resp = requests.post(
            f"{AILANG_PARSE_URL}/api/v1/parse",
            files={"file": open(filepath, "rb")}
        )
        return resp.json()

    elif ext == ".pdf":
        # LlamaParse: cloud LLM-powered PDF parsing
        from llama_parse import LlamaParse
        parser = LlamaParse(result_type="markdown")
        documents = parser.load_data(filepath)
        return {"text": documents[0].text, "metadata": documents[0].metadata}

    else:
        raise ValueError(f"Unsupported format: {ext}")

This routing pattern keeps your existing PDF pipeline intact while adding structural Office parsing at zero marginal cost.

Claude Code Skill

AILANG Parse is available as a Claude Code skill for parsing documents directly within your coding workflow. Install it with a single command:

claude install github:sunholo-data/ailang-parse-skill

Once installed, Claude Code can parse Office documents, extract tables, and convert between formats as part of your development session.

See the Claude Code Skill guide for the full setup walkthrough, usage examples, and supported commands.

REST API

AILANG Parse exposes a full REST API for programmatic access. The API is agent-discoverable — every endpoint is described in the capability manifest with typed errors, cost metadata, and replayability.

Key Endpoints

# Health check
curl https://docparse.ailang.sunholo.com/api/v1/health

# List supported formats
curl https://docparse.ailang.sunholo.com/api/v1/formats

# Parse a document
curl -X POST https://docparse.ailang.sunholo.com/api/v1/parse \
  -F "file=@report.docx"

# Capability manifest (agent discovery)
curl https://docparse.ailang.sunholo.com/api/v1/capabilities

See the API Reference for the full endpoint documentation, authentication, and OpenAPI spec.

Self-Hosting

Prefer containers? Build and run AILANG Parse locally with Docker for batch CLI processing:

# Build the image
git clone https://github.com/sunholo-data/ailang-parse.git
cd ailang-parse
docker build -t docparse .

# Parse a file (mount your documents into /data)
docker run -v $(pwd):/data docparse /data/report.docx

# Parse with AI (pass your API key)
docker run -e GOOGLE_API_KEY="your-key" \
  -v $(pwd):/data docparse --ai gemini-2.5-flash /data/document.pdf

The image builds AILANG from source and uses a CLI entrypoint for direct parsing. For REST API access, use the hosted API with Python, JS, or Go SDKs. See the Run Locally guide for full installation options.