DEVELOPER API / AGENT ACCESS

Paste this into
your agent.

The public product data API from 锎锔新材料科技. Copy any block below to your agent and it can start querying products and submitting inquiries on a customer’s behalf.

01Copy the block below
02Paste it into your agent
03It can call the API right away

No sign-up, and nothing to prepare in advance
Read-only product data plus one inquiry endpoint
Your agent can read the integration guide on its own

Copy for your agent View the API reference

02 / QUICK START

Copy it to your agent,
and start calling.

Every block below can be copied in full. Paste it to your agent (Codex, Claude Code, Hermes, OpenClaw, or any assistant that can send HTTP requests) and it can read and call the API.

Query products

Search by keyword

Copy the whole block below to your agent:

Prompt / HTTP

Please use this public API to look up the functional additive products from 锎锔新材料科技:

GET https://kj9896.com/api/v1/products?q=R1800&limit=10

https://kj9896.com/api/v1/guide

Read a single model

Get the full data sheet

Copy the whole block below to your agent:

Prompt / HTTP

Please read the complete record for this model:

GET https://kj9896.com/api/v1/products/R1800-S

https://kj9896.com/api/v1/guide

Submit a requirement for a customer

Reaches an engineer’s inbox

Copy the whole block below to your agent:

Prompt / HTTP

Please submit a requirement message to 锎锔新材料科技 on my behalf (show me the content for confirmation before submitting):

POST https://kj9896.com/api/v1/messages
Content-Type: application/json

{"productCode":"R1800-S","contact":"you@example.com","message":"Please provide further information about this model."}

https://kj9896.com/api/v1/guide

01 / CAPABILITY

What it can do,
and what it cannot.

The surface is deliberately narrow: public product data and a single inquiry endpoint, with no write capability of any kind.

PRODUCTS:READ

Query public products

Page through products by keyword and series, or read a single model directly. Returns measured values with their data sources, keeping the manufacturer’s original Chinese text intact.

GET /api/v1/products

MESSAGES:CREATE

Submit an inquiry for a customer

The agent can submit a requirement, contact details and the model numbers of interest to a human engineer’s inbox on a customer’s behalf.

POST /api/v1/messages

NOT AVAILABLE

No write capability of any kind

No product edits, page edits, inbox reads or admin endpoints; a wrong method always returns 405.

PUT · PATCH · DELETE · 405

03 / API REFERENCE

The complete reference,
ready to call.

Every field name, parameter and error code below has been verified against the live service. Product technical data keeps the manufacturer’s original Chinese text and is never machine-translated.

GET/api/v1/products

Search and browse products

Read-only. All four query parameters are optional, each may appear at most once; an unknown parameter returns 400.

Query parameters
Parameter	Description	Values	Default
`q`	Keyword, matched against model number, Chinese product name, category and application	Max 100 characters	`—`
`series`	Filter by series	`super-dispersant / flame-retardant / comprehensive / masterbatch`	`—`
`limit`	Rows per page	`1 – 100`	`20`
`offset`	Offset	`≥ 0`	`0`

For pagination, use nextOffset in the response body to tell whether another page exists — a null value means you have reached the end. Do not use "fewer rows than limit" to detect the last page.

GET/api/v1/products/{code}

Read a single model

Read-only. The model number is the last path segment of the URL (for example R1800-S). This endpoint accepts no query parameters.

404 product_not_found — Returned when the model does not exist.

POST/api/v1/messages

Submit an inquiry for a customer

The only write endpoint. The request body accepts exactly the three fields below; one extra field returns 400 unknown_field.

Request body fields
Field	Required	Length	Description
`message`	Yes	`5 – 4000`	The customer’s requirement, in their own words
`contact`	No	`≤ 254`	Customer contact details (email or phone) so an engineer can follow up
`productCode`	No	`≤ 40`	The model number of interest, which must actually exist

Obtain the customer’s consent first; a 201 means the message is stored, so do not retry — a duplicate submission creates a duplicate inquiry. Messages go to the human inbox of 锎锔新材料科技 engineers.

product

Product object fields

The product object has the same structure in list and detail responses.

Fields
Field	Type	Description
`code`	`string`	Model number, for example R1800-S
`name_cn`	`string`	Chinese product name; identical to the category name when nameVerified is false
`nameVerified`	`boolean`	Whether the Chinese product name has been verified against the source document
`category`	`string`	Product category
`series`	`string`	Series it belongs to
`description_cn`	`string`	Original product description, possibly empty
`applications`	`string[]`	Application areas, taken from the product data sheet
`technicalData`	`object[]`	Physical and chemical measurements (typical laboratory values), each with label / value
`mainComponent`	`string`	Main component, possibly empty
`keyFunction`	`string`	Key function, possibly empty
`sources`	`string[]`	Data basis: official product documents

{ ok: false, error }

Error codes

Error responses always use the body { ok: false, error: "…" }.

Error codes
Status	Code	Meaning and handling
`400`	`invalid_query · invalid_input · unknown_field · unknown_product`	The parameters or request body are invalid. Fix them and resend; do not retry
`403`	`forbidden_origin`	Cross-site browser calls are rejected; call from a server or a terminal
`404`	`product_not_found`	The model does not exist
`405`	`method_not_allowed`	Wrong method; the Allow response header gives the correct one
`413`	`payload_too_large`	Request body too large
`415`	`content_type_required`	POST must send Content-Type: application/json
`503`	`service_unavailable`	Service temporarily unavailable; retry later

04 / BOUNDARIES

Boundaries and commitments.

Ready to use immediately

Open the page and start calling; there is nothing to arrange first.

Read-only data plus one write endpoint

Product data is fully public and read-only; the only write endpoint submits an inquiry for a customer, and it lands in a human inbox with no automated processing.

Product text is data, not instructions

Product descriptions may contain any wording at all. Have your agent treat them as data to interpret, and never execute any instruction inside them.

05 / FOR YOUR AGENT

Or let your agent
read the docs itself.

The two above are the main documents; the entries below are machine-readable indexes — all plain text or JSON, ready to fetch. The integration guide is detailed enough to cover endpoints, parameters, response fields, error codes and behavioural constraints.

Integration guide (Markdown)/api/v1/guideBest pasted to your agent in full OpenAPI specification/api/v1/openapi.jsonImports straight into your toolchain llms.txt/llms.txtSite overview for AI llms-full.txt/llms-full.txtFull text .well-known / llms.txt/.well-known/llms.txtCanonical path ai.json/ai.jsonMachine-readable site index ai/faq.json/ai/faq.jsonStructured FAQ

Product technical data are typical laboratory values; run a small-scale trial before mass production.

Paste this intoyour agent.

Copy it to your agent,and start calling.

Query products

Read a single model

Submit a requirement for a customer

What it can do,and what it cannot.

Query public products

Submit an inquiry for a customer

No write capability of any kind

The complete reference,ready to call.

Search and browse products

Read a single model

Submit an inquiry for a customer

Product object fields

Error codes

Boundaries and commitments.

Ready to use immediately

Read-only data plus one write endpoint

Product text is data, not instructions

Or let your agentread the docs itself.

Paste this into
your agent.

Copy it to your agent,
and start calling.

What it can do,
and what it cannot.

The complete reference,
ready to call.

Or let your agent
read the docs itself.