supplygraph-ai

Due Diligence Agent

Overview

The Due Diligence Agent generates a comprehensive, structured, and continually updated intelligence report for any company worldwide.

It consolidates fragmented public data — including corporate registration, legal records, subsidiaries, IP assets, compliance, financial disclosures, workforce, and regulatory signals — into a standardized, enterprise-grade due diligence dossier.

This agent is designed for:

M&A screening
Supplier onboarding
Investment research
Risk and compliance reviews
Strategic partner evaluation

Pain

Traditional due diligence is slow, fragmented, and expensive.

Analysts often spend 10+ hours per company collecting unstructured information from dozens of sources — and by the time the report is finished, much of the data is already outdated.

Key challenges include:

Manual data aggregation
Inconsistent formats
Incomplete coverage
Lack of continuous monitoring
High research costs

This creates blind spots in strategic decision-making and risk management.

Breakthrough

The Due Diligence Agent automates a process that previously required multiple teams and tools.

Behind the scenes, it:

Aggregates data from a global corporate graph
Resolves entities and normalizes records
Structures information into standardized modules
Continuously refreshes and updates changes

For users, the workflow is simple:

Provide a company name
Confirm the matched entity
Receive a complete or chapter-specific due diligence report

Results are:

Structured
Auditable
Enterprise-ready
Continuously updated

Why SupplyGraph AI

100M+ global companies
8,000+ industry benchmarks
1M+ key products and entities

SupplyGraph AI delivers institution-grade due diligence intelligence without requiring you to upload any proprietary internal data.

Every insight is:

Evidence-linked
Contextualized
Continuously monitored
Built for compliance and risk teams

Try the Due Diligence Agent (Live Chatbot)

Before integrating via API, you can experience this agent directly through our interactive chatbot.

This live demo allows you to:

Enter a company name
Confirm the correct legal entity
Generate a full or chapter-specific report
Preview the structure and quality of the final output

Launch the Due Diligence Chatbot
https://supplygraph.ai/zk_chat_os/agentic/dialog.html?name=due_diligence_report

To use the chatbot, you’ll first need to:

Create a SupplyGraph AI account
Top up your credit balance

The chatbot is powered by the same Due Diligence Agent and A2A endpoints described in this documentation.
Credits consumed in the chatbot are deducted in the same way as API / A2A usage.

Everything you experience here can be fully embedded into your own system through A2A integration.

Sandbox Key Support (for Development)

This agent supports Sandbox API Keys, enabling developers to test integrations without consuming credits.

When using a Sandbox Key:

No credits are deducted
The agent returns predefined real sample data for companies such as Tesla, Inc. or BYD, depending on your input
Returned data is real and structured, but limited to the predefined sample set
Processing is accelerated, and results are returned instantly
Only input validation and structural checks are performed

Sandbox Keys are ideal for:

UI development
Validating your processing logic against realistic datasets
SDK / API behavior testing

⚠️ Sandbox results are not dynamically generated — they come from a predefined dataset and must not be used in production analytics.

For a full comparison of Production vs. Sandbox keys, see
Getting Started Guide → API Keys.

Input Requirements

agent_id: due_diligence_report · MCP tool: due_diligence_report

Pricing is per chapter — see Chapter pricing below. Full report (ALL): 26,400 credits.

Input differs by integration surface:

A2A & Agent API

Field	Required	Description
`text`	Yes	Company name with country/region (e.g. `"Tesla, Inc. United States"`)
`chapter_name`	No	Report section to generate; default `ALL`

Example: "Tesla, Inc. United States" with optional "chapter_name": "Company Registration Information"

The agent resolves the company name and may ask you to confirm the matched entity before report generation starts (see Multi-turn Example (A2A)).

MCP

MCP does not use multi-turn company confirmation on due_diligence_report. Resolve pid first, then pass pid and chapter_name.

Step	MCP tool	Input	Output
1	`search_company_candidates`	`text`	`candidates[].pid`
2	`due_diligence_report`	`pid`, `chapter_name`	Due diligence report

Step 1 example: {"text": "Tesla United States"}
Step 2 example: {"pid": "a77828f060c866441f2403384b271e63", "chapter_name": "ALL"}

search_company_candidates pricing: 1 credit / run. See MCP tool config below.

Full machine-readable schemas → GET /api/v1/agents/due_diligence_report/manifest (see Agent API §3).

Chapter pricing

`chapter_name`	Credits / run
`ALL` (full report)	26,400
`Company Registration Information`	1,518
`Branch Offices`	1,650
`Patent Holdings`	12,045
`Shareholder Structure`	1,386
…	…

42 chapters total. See agent manifest → pricing.options for the complete list and current rates.

Output Format

Primary output lives in data.content (Agent API) or task artifacts (A2A / MCP). The payload is a oneOf:

State	`content` type	Description
In progress, failed, cancelled, or waiting for user input	`string`	Text or Markdown — validation prompts, confirmation requests, or error messages
Task completed successfully	`object`	Structured due diligence report (see below)

Structured success object:

{
  "type": "results",
  "data": {
    "languages": ["en", "zk"],
    "report_infos": {
      "en": {
        "report_date": "2025-11-30",
        "report_name": "Due Diligence Report — Tesla, Inc.",
        "report_type": "dd report"
      },
      "zh": {
        "report_date": "2025-11-30",
        "report_name": "尽职调查报告 — Tesla, Inc.",
        "report_type": "dd report"
      }
    },
    "chapter_infos": [
      {
        "en": {
          "chapter_name": "Company Registration Information",
          "sub_sections": [],
          "chapter_texts": ["Markdown paragraph…"]
        },
        "zh": {
          "chapter_name": "工商注册信息",
          "sub_sections": [],
          "chapter_texts": ["Markdown 段落…"]
        }
      }
    ]
  }
}

Field	Description
`type`	Always `"results"` on success
`data.languages`	Language codes in the response — `en` (English), `zk` (Chinese)
`data.report_infos`	Report metadata keyed by language (`en` required; `zh` when Chinese is included)
`data.chapter_infos[]`	Array of chapters, each with `en` / `zh` blocks
`chapter_infos[].en\\|zh.chapter_name`	Chapter title
`chapter_infos[].en\\|zh.sub_sections[]`	Nested sub-sections (`section_title`, `sub_sections`, `section_texts`)
`chapter_infos[].en\\|zh.chapter_texts[]`	Chapter body paragraphs (Markdown)

Estimated task duration: 1–10 hours (Production). Sandbox returns instantly.

After entity confirmation (A2A) or pid submission (MCP), poll status until TASK_COMPLETED, then call results.

Sample Response (Sandbox)

Sandbox returns Tesla, Inc. or BYD fixture data with the same structure as Production. Content below is abbreviated.

Success (TASK_COMPLETED) — Agent API results:

{
  "success": true,
  "code": "TASK_COMPLETED",
  "message": "Task completed successfully.",
  "data": {
    "task_id": "<task-id>",
    "agent": "due_diligence_report",
    "stage": "completed",
    "progress": 100,
    "content": {
      "type": "results",
      "data": {
        "languages": ["en", "zk"],
        "report_infos": {
          "en": {
            "report_date": "2025-11-30",
            "report_name": "Due Diligence Report — Tesla, Inc.",
            "report_type": "dd report"
          },
          "zh": {
            "report_date": "2025-11-30",
            "report_name": "尽职调查报告 — Tesla, Inc.",
            "report_type": "dd report"
          }
        },
        "chapter_infos": [
          {
            "en": {
              "chapter_name": "Company Registration Information",
              "sub_sections": [
                {
                  "section_title": "Basic Registration",
                  "sub_sections": [],
                  "section_texts": ["Tesla, Inc. is registered in the State of Delaware, United States."]
                }
              ],
              "chapter_texts": ["### Overview\nCorporate registration records confirm active standing."]
            },
            "zh": {
              "chapter_name": "工商注册信息",
              "sub_sections": [],
              "chapter_texts": ["### 概述\n工商登记信息显示企业处于存续状态。"]
            }
          }
        ]
      }
    }
  },
  "metadata": { "credits_used": 0 },
  "errors": null
}

Multi-turn Example (A2A)

A2A and Agent API only. MCP uses search_company_candidates → due_diligence_report instead of multi-turn confirmation.

When the agent cannot unambiguously match a company from text, it returns WAITING_USER with a Markdown prompt in content (string). Continue with the same task_id.

Turn 1 — submit company name (optional chapter_name in body or via manifest defaults):

curl -X POST "https://agent.supplygraph.ai/api/v1/agents/due_diligence_report/run" \
  -H "Authorization: Bearer <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{"text": "Tesla, Inc. United States", "chapter_name": "ALL", "stream": false}'

Poll until WAITING_USER. Example content:

{
  "success": true,
  "code": "WAITING_USER",
  "data": {
    "task_id": "<task-id>",
    "content": "Here are the companies we've identified.\nCompany Name: Tesla, Inc.\nCountry: United States\nPlease reply [Yes] or [No] to confirm."
  }
}

Turn 2 — confirm entity (same task_id):

curl -X POST "https://agent.supplygraph.ai/api/v1/agents/due_diligence_report/run" \
  -H "Authorization: Bearer <YOUR_API_KEY>" \
  -H "Content-Type: application/json" \
  -d '{"text": "yes", "task_id": "<task-id>", "stream": false}'

The agent queues report generation. Poll status until TASK_COMPLETED (may take 1–10 h in Production), then retrieve structured results via results.

See Agent API §8 for the general multi-turn pattern.

MCP Resolution Workflow

Step 1 — `search_company_candidates`

{
  "type": "company_candidate_search",
  "data": {
    "candidates": [
      {
        "pid": "a77828f060c866441f2403384b271e63",
        "company_name": "Tesla, Inc.",
        "country_name": "United States"
      }
    ]
  }
}

Step 2 — `due_diligence_report`

report = await session.experimental.call_tool_as_task(
    "due_diligence_report",
    {
        "pid": "a77828f060c866441f2403384b271e63",
        "chapter_name": "Company Registration Information",  # or "ALL"
    },
)
# poll until complete; parse type: "results"

MCP Tool Config

MCP input_schema for due_diligence_report:

{
  "enabled": true,
  "input_schema": {
    "type": "object",
    "additionalProperties": false,
    "properties": {
      "pid": {
        "type": "string",
        "description": "Internal company ID for the target enterprise, obtained from the search_company_candidates MCP tool (e.g. a77828f060c866441f2403384b271e63 for Tesla, Inc.)."
      },
      "chapter_name": {
        "type": "string",
        "description": "Report chapter to generate. Use ALL for the full due diligence report, or a specific chapter name from the agent manifest.",
        "enum": [
          "ALL",
          "Company Registration Information",
          "Branch Offices",
          "Corporate Brand Initiatives",
          "Administrative Sanctions",
          "Software Copyright Details",
          "Outbound Investments",
          "Financing Activities",
          "Competitor Analysis",
          "Subsidiary Companies",
          "Trademark Portfolio",
          "Patent Holdings",
          "Website Registrations",
          "Court Judgments",
          "Shareholder Structure",
          "Senior Management Team",
          "Administrative Permits",
          "Court Hearing Notices",
          "Court Notices",
          "Equity Pledges",
          "Mobile Applications",
          "Copyrighted Works",
          "Equity Freezes",
          "Chattel Mortgages",
          "WeChat Official Accounts",
          "Tendering and Bidding Activities",
          "Qualification Certificates",
          "Engineering Irregularities",
          "Major Regulatory Violations",
          "Compensation and Benefits",
          "Enforcement Targets",
          "Supplier Network",
          "Credit Ratings",
          "Tax Offenses",
          "Regulatory Spot Checks",
          "Import-Export Credit Records",
          "Regulatory Actions",
          "Granted Government Subsidies",
          "Eligible Government Subsidies",
          "Consolidated Statements of Operations",
          "Income Statement",
          "Statement of Cash Flows",
          "Consolidated Balance Sheets"
        ],
        "default": "ALL"
      }
    },
    "required": ["pid"]
  },
  "execution": {
    "taskSupport": "required"
  },
  "runtime": {
    "poll_interval_seconds": 30,
    "timeout_seconds": 600,
    "task_ttl_ms": 36000000
  }
}

chapter_name is optional; omit to generate the full report (ALL). Production reports may run 1–10 h — consider increasing task_ttl_ms accordingly.

Integration

Method	ID / Tool	Documentation
A2A	`due_diligence_report`	a2a.md
MCP	`search_company_candidates` + `due_diligence_report`	mcp.md
Agent API	`due_diligence_report`	agent-api.md

Call pattern is identical across agents — only agent_id / tool name and input differ.

Quick examples: A2A / MCP · Agent API

Errors

Common codes → Agent API §10.

Situation	Code	Agent behavior
Ambiguous or unmatched company (A2A / Agent API)	`WAITING_USER`	Prompts for entity confirmation; resume with same `task_id`
Invalid `chapter_name`	`WAITING_USER`	Returns guidance to choose a valid chapter from the enum
Invalid `pid` (MCP)	`TASK_FAILED`	Returns guidance in `content` (string) — re-run `search_company_candidates`
Insufficient credits	`INSUFFICIENT_CREDITS`	Top up via Console
Report still generating	`TASK_RUNNING`	Continue polling `status`; Production may take up to 10 hours

This site is open source. Improve this page.