@pdfvector/instance-contract 0.0.19 → 0.0.21

package/.tsc/lib/pdfvector-model-schema.js

@@ -1,2 +1,4 @@
 import { z } from "zod";
-export const pdfvectorModelSchema = z.enum(["nano", "mini", "pro", "max"]);
+export const pdfvectorModelSchema = z.enum(["nano", "mini", "pro", "max"], {
+    message: "model must be one of: nano, mini, pro, max",
+});

package/.tsc/lib/router/academic/fetch.d.ts

@@ -0,0 +1,46 @@
+import { z } from "zod";
+export declare const fetch: import("@orpc/contract").ContractProcedureBuilderWithInputOutput<z.ZodObject<{
+    ids: z.ZodArray<z.ZodString>;
+    fields: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodEnum<{
+        url: "url";
+        doi: "doi";
+        title: "title";
+        providerURL: "providerURL";
+        date: "date";
+        authors: "authors";
+        year: "year";
+        totalCitations: "totalCitations";
+        totalReferences: "totalReferences";
+        abstract: "abstract";
+        pdfURL: "pdfURL";
+        provider: "provider";
+        providerData: "providerData";
+    }>>>>;
+}, z.core.$strip>, z.ZodObject<{
+    results: z.ZodArray<z.ZodObject<{
+        id: z.ZodString;
+        detectedProvider: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        doi: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        title: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        url: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        providerURL: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        authors: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodObject<{
+            name: z.ZodString;
+            url: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        }, z.core.$strip>>>>;
+        date: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        year: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        totalCitations: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        totalReferences: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        abstract: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        pdfURL: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        provider: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        providerData: z.ZodOptional<z.ZodNullable<z.ZodRecord<z.ZodString, z.ZodUnknown>>>;
+    }, z.core.$strip>>;
+    errors: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodObject<{
+        id: z.ZodString;
+        error: z.ZodString;
+        code: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+    }, z.core.$strip>>>>;
+    requestId: z.ZodNumber;
+}, z.core.$strip>, Record<never, never>, Record<never, never>>;

package/.tsc/lib/router/academic/fetch.js

@@ -0,0 +1,101 @@
+import { oc } from "@orpc/contract";
+import { z } from "zod";
+const authorSchema = z.object({
+    name: z.string(),
+    url: z.string().nullish(),
+});
+const publicationSchema = z.object({
+    id: z.string().describe("Original input identifier"),
+    detectedProvider: z
+        .string()
+        .nullish()
+        .describe("Provider that returned this result"),
+    doi: z.string().nullish(),
+    title: z.string().nullish(),
+    url: z.string().nullish(),
+    providerURL: z.string().nullish(),
+    authors: z.array(authorSchema).nullish(),
+    date: z.string().nullish(),
+    year: z.number().nullish(),
+    totalCitations: z.number().nullish(),
+    totalReferences: z.number().nullish(),
+    abstract: z.string().nullish(),
+    pdfURL: z.string().nullish(),
+    provider: z.string().nullish(),
+    providerData: z.record(z.string(), z.unknown()).nullish(),
+});
+const fetchErrorSchema = z.object({
+    id: z.string(),
+    error: z.string(),
+    code: z.string().nullish(),
+});
+const requestExamples = {
+    "Fetch by DOI": {
+        summary: "Fetch a paper by DOI",
+        value: {
+            ids: ["10.1038/nature12373"],
+        },
+    },
+    "Fetch multiple IDs": {
+        summary: "Fetch papers by mixed identifier types",
+        value: {
+            ids: ["10.1038/nature12373", "2301.08745", "33293753"],
+        },
+    },
+    "Fetch ArXiv paper": {
+        summary: "Fetch by ArXiv ID",
+        value: {
+            ids: ["2303.08774"],
+        },
+    },
+};
+export const fetch = oc
+    .input(z.object({
+    ids: z
+        .array(z.string({ message: "Each ID must be a string" }), {
+        message: "ids is required and must be an array of strings",
+    })
+        .min(1, "ids must contain at least 1 identifier")
+        .max(100, "ids must contain at most 100 identifiers")
+        .describe("Publication identifiers to fetch. Supports DOI, PubMed ID, ArXiv ID, Semantic Scholar ID, ERIC ID, Europe PMC ID, OpenAlex ID."),
+    fields: z
+        .array(z.enum([
+        "doi",
+        "title",
+        "url",
+        "providerURL",
+        "authors",
+        "date",
+        "year",
+        "totalCitations",
+        "totalReferences",
+        "abstract",
+        "pdfURL",
+        "provider",
+        "providerData",
+    ]), {
+        message: "fields must be an array of valid publication field names",
+    })
+        .nullish()
+        .describe("Specific publication fields to return. If omitted, returns all fields except providerData."),
+}))
+    .output(z.object({
+    results: z.array(publicationSchema),
+    errors: z.array(fetchErrorSchema).nullish(),
+    requestId: z.number().int(),
+}))
+    .route({
+    summary: "Fetch academic papers by ID",
+    description: "Fetch specific academic publications by their unique identifiers with automatic provider detection and batch processing.",
+    tags: ["Academic"],
+    spec: (op) => {
+        op.security = [{ bearerAuth: [] }];
+        const reqBody = op.requestBody;
+        if (reqBody?.content) {
+            for (const mediaType of Object.values(reqBody.content)) {
+                mediaType.examples = requestExamples;
+            }
+        }
+        return op;
+    },
+});

package/.tsc/lib/router/academic/find-citations.d.ts

@@ -0,0 +1,55 @@
+import { z } from "zod";
+export declare const findCitations: import("@orpc/contract").ContractProcedureBuilderWithInputOutput<z.ZodObject<{
+    paragraph: z.ZodString;
+    providers: z.ZodDefault<z.ZodArray<z.ZodEnum<{
+        "semantic-scholar": "semantic-scholar";
+        pubmed: "pubmed";
+        arxiv: "arxiv";
+        "google-scholar": "google-scholar";
+        eric: "eric";
+        "europe-pmc": "europe-pmc";
+        openalex: "openalex";
+    }>>>;
+    fields: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodEnum<{
+        url: "url";
+        doi: "doi";
+        title: "title";
+        providerURL: "providerURL";
+        date: "date";
+        authors: "authors";
+        year: "year";
+        totalCitations: "totalCitations";
+        totalReferences: "totalReferences";
+        abstract: "abstract";
+        pdfURL: "pdfURL";
+        provider: "provider";
+        providerData: "providerData";
+    }>>>>;
+}, z.core.$strip>, z.ZodObject<{
+    results: z.ZodArray<z.ZodObject<{
+        sentence: z.ZodString;
+        citations: z.ZodArray<z.ZodObject<{
+            score: z.ZodNumber;
+            doi: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            title: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            url: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            providerURL: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            authors: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodObject<{
+                name: z.ZodString;
+                url: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            }, z.core.$strip>>>>;
+            date: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            year: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+            totalCitations: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+            totalReferences: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+            abstract: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            pdfURL: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            provider: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+            providerData: z.ZodOptional<z.ZodNullable<z.ZodRecord<z.ZodString, z.ZodUnknown>>>;
+        }, z.core.$strip>>;
+    }, z.core.$strip>>;
+    sentenceCount: z.ZodNumber;
+    totalCitations: z.ZodNumber;
+    errors: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodString>>>;
+    requestId: z.ZodNumber;
+}, z.core.$strip>, Record<never, never>, Record<never, never>>;

package/.tsc/lib/router/academic/find-citations.js

@@ -0,0 +1,108 @@
+import { oc } from "@orpc/contract";
+import { z } from "zod";
+const providerSchema = z.enum([
+    "semantic-scholar",
+    "pubmed",
+    "arxiv",
+    "google-scholar",
+    "eric",
+    "europe-pmc",
+    "openalex",
+], {
+    message: "Invalid provider. Must be one of: semantic-scholar, pubmed, arxiv, google-scholar, eric, europe-pmc, openalex",
+});
+const authorSchema = z.object({
+    name: z.string(),
+    url: z.string().nullish(),
+});
+const scoredPublicationSchema = z.object({
+    score: z.number().describe("Relevance score (0-10)"),
+    doi: z.string().nullish(),
+    title: z.string().nullish(),
+    url: z.string().nullish(),
+    providerURL: z.string().nullish(),
+    authors: z.array(authorSchema).nullish(),
+    date: z.string().nullish(),
+    year: z.number().nullish(),
+    totalCitations: z.number().nullish(),
+    totalReferences: z.number().nullish(),
+    abstract: z.string().nullish(),
+    pdfURL: z.string().nullish(),
+    provider: z.string().nullish(),
+    providerData: z.record(z.string(), z.unknown()).nullish(),
+});
+const sentenceCitationSchema = z.object({
+    sentence: z.string(),
+    citations: z.array(scoredPublicationSchema),
+});
+const requestExamples = {
+    "Find citations for a paragraph": {
+        summary: "Find citations for a research paragraph",
+        value: {
+            paragraph: "Deep learning has revolutionized computer vision and natural language processing. Transformer architectures have become the dominant approach for sequence modeling tasks. Recent large language models demonstrate emergent capabilities not seen in smaller models.",
+        },
+    },
+    "Find citations with multiple providers": {
+        summary: "Search across multiple databases",
+        value: {
+            paragraph: "CRISPR-Cas9 gene editing technology has transformed molecular biology. The ability to precisely modify DNA sequences has opened new avenues for treating genetic diseases.",
+            providers: ["semantic-scholar", "pubmed"],
+        },
+    },
+};
+export const findCitations = oc
+    .input(z.object({
+    paragraph: z
+        .string({ message: "paragraph is required and must be a string" })
+        .min(1, "paragraph must not be empty")
+        .max(5000, "paragraph must be at most 5000 characters")
+        .describe("Text paragraph to find relevant academic citations for"),
+    providers: z
+        .array(providerSchema, {
+        message: "providers must be an array of valid provider names",
+    })
+        .default(["semantic-scholar"])
+        .describe("Academic databases to search for citations"),
+    fields: z
+        .array(z.enum([
+        "doi",
+        "title",
+        "url",
+        "providerURL",
+        "authors",
+        "date",
+        "year",
+        "totalCitations",
+        "totalReferences",
+        "abstract",
+        "pdfURL",
+        "provider",
+        "providerData",
+    ]), {
+        message: "fields must be an array of valid publication field names",
+    })
+        .nullish()
+        .describe("Specific citation fields to return. If omitted, returns all fields except providerData."),
+}))
+    .output(z.object({
+    results: z.array(sentenceCitationSchema),
+    sentenceCount: z.number().int(),
+    totalCitations: z.number().int(),
+    errors: z.array(z.string()).nullish(),
+    requestId: z.number().int(),
+}))
+    .route({
+    summary: "Find citations for a paragraph",
+    description: "Find relevant academic citations for a paragraph by splitting it into sentences and scoring papers using semantic embedding similarity.",
+    tags: ["Academic"],
+    spec: (op) => {
+        op.security = [{ bearerAuth: [] }];
+        const reqBody = op.requestBody;
+        if (reqBody?.content) {
+            for (const mediaType of Object.values(reqBody.content)) {
+                mediaType.examples = requestExamples;
+            }
+        }
+        return op;
+    },
+});

package/.tsc/lib/router/academic/index.d.ts

@@ -0,0 +1,3 @@
+export * from "./fetch";
+export * from "./find-citations";
+export * from "./search";

package/.tsc/lib/router/academic/index.js

@@ -0,0 +1,3 @@
+export * from "./fetch";
+export * from "./find-citations";
+export * from "./search";

package/.tsc/lib/router/academic/search.d.ts

@@ -0,0 +1,67 @@
+import { z } from "zod";
+export declare const search: import("@orpc/contract").ContractProcedureBuilderWithInputOutput<z.ZodObject<{
+    query: z.ZodString;
+    providers: z.ZodDefault<z.ZodArray<z.ZodEnum<{
+        "semantic-scholar": "semantic-scholar";
+        pubmed: "pubmed";
+        arxiv: "arxiv";
+        "google-scholar": "google-scholar";
+        eric: "eric";
+        "europe-pmc": "europe-pmc";
+        openalex: "openalex";
+    }>>>;
+    offset: z.ZodDefault<z.ZodNumber>;
+    limit: z.ZodDefault<z.ZodNumber>;
+    yearFrom: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+    yearTo: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+    fields: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodEnum<{
+        url: "url";
+        doi: "doi";
+        title: "title";
+        providerURL: "providerURL";
+        date: "date";
+        authors: "authors";
+        year: "year";
+        totalCitations: "totalCitations";
+        totalReferences: "totalReferences";
+        abstract: "abstract";
+        pdfURL: "pdfURL";
+        provider: "provider";
+        providerData: "providerData";
+    }>>>>;
+}, z.core.$strip>, z.ZodObject<{
+    estimatedTotalResults: z.ZodNumber;
+    results: z.ZodArray<z.ZodObject<{
+        doi: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        title: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        url: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        providerURL: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        authors: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodObject<{
+            name: z.ZodString;
+            url: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        }, z.core.$strip>>>>;
+        date: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        year: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        totalCitations: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        totalReferences: z.ZodOptional<z.ZodNullable<z.ZodNumber>>;
+        abstract: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        pdfURL: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        provider: z.ZodOptional<z.ZodNullable<z.ZodEnum<{
+            "semantic-scholar": "semantic-scholar";
+            pubmed: "pubmed";
+            arxiv: "arxiv";
+            "google-scholar": "google-scholar";
+            eric: "eric";
+            "europe-pmc": "europe-pmc";
+            openalex: "openalex";
+        }>>>;
+        providerData: z.ZodOptional<z.ZodNullable<z.ZodRecord<z.ZodString, z.ZodUnknown>>>;
+    }, z.core.$strip>>;
+    errors: z.ZodOptional<z.ZodNullable<z.ZodArray<z.ZodObject<{
+        provider: z.ZodString;
+        message: z.ZodString;
+        code: z.ZodOptional<z.ZodNullable<z.ZodString>>;
+        details: z.ZodOptional<z.ZodNullable<z.ZodRecord<z.ZodString, z.ZodUnknown>>>;
+    }, z.core.$strip>>>>;
+    requestId: z.ZodNumber;
+}, z.core.$strip>, Record<never, never>, Record<never, never>>;

package/.tsc/lib/router/academic/search.js

@@ -0,0 +1,146 @@
+import { oc } from "@orpc/contract";
+import { z } from "zod";
+const providerSchema = z.enum([
+    "semantic-scholar",
+    "pubmed",
+    "arxiv",
+    "google-scholar",
+    "eric",
+    "europe-pmc",
+    "openalex",
+], {
+    message: "Invalid provider. Must be one of: semantic-scholar, pubmed, arxiv, google-scholar, eric, europe-pmc, openalex",
+});
+const authorSchema = z.object({
+    name: z.string(),
+    url: z.string().nullish(),
+});
+const publicationSchema = z.object({
+    doi: z.string().nullish(),
+    title: z.string().nullish(),
+    url: z.string().nullish(),
+    providerURL: z.string().nullish(),
+    authors: z.array(authorSchema).nullish(),
+    date: z.string().nullish(),
+    year: z.number().nullish(),
+    totalCitations: z.number().nullish(),
+    totalReferences: z.number().nullish(),
+    abstract: z.string().nullish(),
+    pdfURL: z.string().nullish(),
+    provider: providerSchema.nullish(),
+    providerData: z.record(z.string(), z.unknown()).nullish(),
+});
+const providerErrorSchema = z.object({
+    provider: z.string(),
+    message: z.string(),
+    code: z.string().nullish(),
+    details: z.record(z.string(), z.unknown()).nullish(),
+});
+const requestExamples = {
+    "Search with default provider": {
+        summary: "Search with Semantic Scholar",
+        value: {
+            query: "deep learning for natural language processing",
+        },
+    },
+    "Search multiple providers": {
+        summary: "Search across multiple databases",
+        value: {
+            query: "CRISPR gene editing",
+            providers: ["semantic-scholar", "pubmed", "openalex"],
+            limit: 10,
+        },
+    },
+    "Search with year filter": {
+        summary: "Search recent papers",
+        value: {
+            query: "large language models",
+            providers: ["semantic-scholar"],
+            yearFrom: 2023,
+            yearTo: 2026,
+            limit: 20,
+        },
+    },
+};
+export const search = oc
+    .input(z.object({
+    query: z
+        .string({ message: "query is required and must be a string" })
+        .min(1, "query must not be empty")
+        .max(400, "query must be at most 400 characters")
+        .describe("Search query string for academic papers"),
+    providers: z
+        .array(providerSchema, {
+        message: "providers must be an array of valid provider names",
+    })
+        .default(["semantic-scholar"])
+        .describe("Academic databases to search"),
+    offset: z
+        .number({ message: "offset must be a number" })
+        .int("offset must be a whole number")
+        .min(0, "offset must be 0 or greater")
+        .default(0)
+        .describe("Results offset"),
+    limit: z
+        .number({ message: "limit must be a number" })
+        .int("limit must be a whole number")
+        .min(1, "limit must be at least 1")
+        .max(100, "limit must be at most 100")
+        .default(20)
+        .describe("Results per provider"),
+    yearFrom: z
+        .number({ message: "yearFrom must be a number" })
+        .int("yearFrom must be a whole number")
+        .min(1900, "yearFrom must be 1900 or later")
+        .max(2100, "yearFrom must be 2100 or earlier")
+        .nullish()
+        .describe("Filter papers published from this year"),
+    yearTo: z
+        .number({ message: "yearTo must be a number" })
+        .int("yearTo must be a whole number")
+        .min(1900, "yearTo must be 1900 or later")
+        .max(2100, "yearTo must be 2100 or earlier")
+        .nullish()
+        .describe("Filter papers published up to this year"),
+    fields: z
+        .array(z.enum([
+        "doi",
+        "title",
+        "url",
+        "providerURL",
+        "authors",
+        "date",
+        "year",
+        "totalCitations",
+        "totalReferences",
+        "abstract",
+        "pdfURL",
+        "provider",
+        "providerData",
+    ]), {
+        message: "fields must be an array of valid publication field names",
+    })
+        .nullish()
+        .describe("Specific publication fields to return. If omitted, returns all fields except providerData."),
+}))
+    .output(z.object({
+    estimatedTotalResults: z.number(),
+    results: z.array(publicationSchema),
+    errors: z.array(providerErrorSchema).nullish(),
+    requestId: z.number().int(),
+}))
+    .route({
+    summary: "Search academic papers",
+    description: "Search for academic publications across multiple databases with relevance ranking.",
+    tags: ["Academic"],
+    spec: (op) => {
+        op.security = [{ bearerAuth: [] }];
+        const reqBody = op.requestBody;
+        if (reqBody?.content) {
+            for (const mediaType of Object.values(reqBody.content)) {
+                mediaType.examples = requestExamples;
+            }
+        }
+        return op;
+    },
+});

package/.tsc/lib/router/bankStatement/ask.js

@@ -29,7 +29,7 @@ const askInputSchema = z.object({
         .describe("Base64-encoded bank statement file content"),
     question: z
         .string()
-        .min(4)
+        .min(4, "question must be at least 4 characters")
         .describe("The question to answer about the bank statement"),
     model: specializedModelSchema.describe("Model tier for answering the question. " +
         "'auto' (default): Automatically selects the best tier. " +

package/.tsc/lib/router/bankStatement/extract.js

@@ -29,7 +29,7 @@ const extractInputSchema = z.object({
         .describe("Base64-encoded bank statement file content"),
     prompt: z
         .string()
-        .min(4)
+        .min(4, "prompt must be at least 4 characters")
         .describe("The prompt instructing the AI how to extract data from the bank statement"),
     schema: z
         .union([

package/.tsc/lib/router/bankStatement/parse.js

@@ -2,7 +2,9 @@ import { oc } from "@orpc/contract";
 import { z } from "zod";
 import { getDefaultSpec } from "./get-default-spec";
 const specializedParseModelSchema = z
-    .enum(["pro", "max", "auto"])
+    .enum(["pro", "max", "auto"], {
+    message: "model must be one of: pro, max, auto",
+})
     .default("auto");
 const parseInputSchema = z.object({
     url: z

package/.tsc/lib/router/document/ask.js

@@ -26,7 +26,7 @@ const askInputSchema = z.object({
         .describe("Base64-encoded document file content"),
     question: z
         .string()
-        .min(4)
+        .min(4, "question must be at least 4 characters")
         .describe("The question to answer about the document"),
     model: z
         .enum(["auto", ...pdfvectorModelSchema.options])

package/.tsc/lib/router/document/extract.js

@@ -26,7 +26,7 @@ const extractInputSchema = z.object({
         .describe("Base64-encoded document file content"),
     prompt: z
         .string()
-        .min(4)
+        .min(4, "prompt must be at least 4 characters")
         .describe("The prompt instructing the AI how to extract data from the document"),
     schema: z
         .union([

package/.tsc/lib/router/free/bank-statement-parse.d.ts

@@ -0,0 +1,15 @@
+import { z } from "zod";
+export declare const bankStatementParse: import("@orpc/contract").ContractProcedureBuilderWithInputOutput<z.ZodObject<{
+    url: z.ZodOptional<z.ZodURL>;
+    file: z.ZodOptional<z.ZodFile>;
+    base64: z.ZodOptional<z.ZodString>;
+}, z.core.$strip>, z.ZodObject<{
+    markdown: z.ZodString;
+    pageCount: z.ZodNumber;
+    model: z.ZodEnum<{
+        pro: "pro";
+        max: "max";
+    }>;
+    requestId: z.ZodNumber;
+    html: z.ZodOptional<z.ZodString>;
+}, z.core.$strip>, Record<never, never>, Record<never, never>>;

package/.tsc/lib/router/free/bank-statement-parse.js

@@ -0,0 +1,94 @@
+import { oc } from "@orpc/contract";
+import { z } from "zod";
+const inputSchema = z.object({
+    url: z
+        .url()
+        .optional()
+        .describe("URL of the bank statement file to fetch and parse"),
+    file: z
+        .file()
+        .mime([
+        "application/pdf",
+        "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+        "application/vnd.openxmlformats-officedocument.spreadsheetml.sheet",
+        "text/csv",
+        "application/csv",
+        "image/png",
+        "image/jpeg",
+    ])
+        .optional()
+        .describe("Bank statement file upload via multipart form-data (PDF, DOCX, XLSX, CSV, PNG, JPG)"),
+    base64: z
+        .string()
+        .optional()
+        .describe("Base64-encoded bank statement file content"),
+});
+const outputSchema = z.object({
+    markdown: z
+        .string()
+        .describe("Extracted text content from the bank statement"),
+    pageCount: z.number().int().describe("Total number of pages in the document"),
+    model: z
+        .enum(["pro", "max"])
+        .describe("Model tier used to parse the bank statement"),
+    requestId: z
+        .number()
+        .int()
+        .describe("Unique request identifier for this API call"),
+    html: z
+        .string()
+        .optional()
+        .describe("Full HTML representation of the document content"),
+});
+const requestExamples = {
+    "Parse from URL": {
+        summary: "Parse from URL",
+        value: {
+            url: "https://raw.githubusercontent.com/aws-samples/textract-bank-statement-processor/main/Statement20230428.pdf",
+        },
+    },
+    "Parse from base64": {
+        summary: "Parse from base64",
+        value: {
+            base64: "JVBERi0xLjAKMSAwIG9iajw8L1R5cGUvQ2F0YWxvZy9QYWdlcyAyIDAgUj4+ZW5kb2JqIDIgMCBvYmo8PC9UeXBlL1BhZ2VzL0tpZHNbMyAwIFJdL0NvdW50IDE+PmVuZG9iaiAzIDAgb2JqPDwvVHlwZS9QYWdlL01lZGlhQm94WzAgMCAzIDNdL1BhcmVudCAyIDAgUj4+ZW5kb2JqCnhyZWYKMCA0CjAwMDAwMDAwMDAgNjU1MzUgZiAKMDAwMDAwMDAwOSAwMDAwMCBuIAowMDAwMDAwMDU4IDAwMDAwIG4gCjAwMDAwMDAxMTUgMDAwMDAgbiAKdHJhaWxlcjw8L1NpemUgNC9Sb290IDEgMCBSPj4Kc3RhcnR4cmVmCjE5MAolJUVPRg==",
+        },
+    },
+    "Parse from file upload": {
+        summary: "Parse from file upload",
+        value: {
+            file: "(binary)",
+        },
+    },
+};
+const responseExample = {
+    markdown: "## Bank Statement\n\n### Account Information\n\n| Field | Value |\n|-------|-------|\n| Bank Name | Example Bank |\n| Account Number | ****1234 |\n| Statement Start Date | 2023-03-01 |\n| Statement End Date | 2023-03-31 |\n| Beginning Balance | $5,432.10 |\n| Ending Balance | $4,891.55 |\n\n### Transactions\n\n| Date | Description | Deposit Amount | Withdrawal Amount |\n| --- | --- | --- | --- |\n| 2023-03-05 | Direct Deposit - Payroll | $3,200.00 | |\n| 2023-03-08 | Online Transfer Out | | $500.00 |\n| 2023-03-15 | Grocery Store | | $142.35 |",
+    pageCount: 2,
+    model: "max",
+    requestId: 42,
+    html: "<p>Full HTML content...</p>",
+};
+export const bankStatementParse = oc
+    .route({
+    summary: "Free: Parse a bank statement",
+    description: "Extract text and structured data from a bank statement. Supports PDF, Word, Excel, CSV, and Image files. Free tier: max 5 pages, max 5 MB. Rate limited to 5 requests per IP per day.",
+    tags: ["Free"],
+    spec: (op) => {
+        op.security = [];
+        const reqBody = op.requestBody;
+        if (reqBody?.content) {
+            for (const mediaType of Object.values(reqBody.content)) {
+                mediaType.examples = requestExamples;
+            }
+        }
+        const responses = op.responses;
+        const okResponse = responses?.["200"];
+        if (okResponse?.content) {
+            for (const mediaType of Object.values(okResponse.content)) {
+                mediaType.example = responseExample;
+            }
+        }
+        return op;
+    },
+})
+    .input(inputSchema)
+    .output(outputSchema);

package/.tsc/lib/router/free/generate-schema.d.ts

@@ -0,0 +1,7 @@
+import { z } from "zod";
+export declare const generateSchema: import("@orpc/contract").ContractProcedureBuilderWithInputOutput<z.ZodObject<{
+    instruction: z.ZodString;
+}, z.core.$strip>, z.ZodObject<{
+    schema: z.ZodRecord<z.ZodString, z.ZodUnknown>;
+    requestId: z.ZodNumber;
+}, z.core.$strip>, Record<never, never>, Record<never, never>>;

package/.tsc/lib/router/free/generate-schema.js

@@ -0,0 +1,110 @@
+import { oc } from "@orpc/contract";
+import { z } from "zod";
+const inputSchema = z.object({
+    instruction: z
+        .string()
+        .min(10, "Instruction must be at least 10 characters")
+        .max(2000, "Instruction must be at most 2000 characters")
+        .describe("Describe the data you want to extract, e.g. 'Extract vendor name, total amount, and line items from an invoice'"),
+});
+const outputSchema = z.object({
+    schema: z
+        .record(z.string(), z.unknown())
+        .describe("Generated JSON Schema for structured data extraction"),
+    requestId: z
+        .number()
+        .int()
+        .describe("Unique request identifier for this API call"),
+});
+const requestExamples = {
+    "Invoice extraction": {
+        summary: "Generate schema for invoice extraction",
+        value: {
+            instruction: "Extract vendor name, total amount, invoice date, and line items from an invoice",
+        },
+    },
+    "Resume parsing": {
+        summary: "Generate schema for resume parsing",
+        value: {
+            instruction: "Extract name, email, phone number, work experience, and education from a resume",
+        },
+    },
+    "Receipt data": {
+        summary: "Generate schema for receipt data",
+        value: {
+            instruction: "Extract store name, date, items purchased with prices, subtotal, tax, and total from a receipt",
+        },
+    },
+};
+const responseExample = {
+    schema: {
+        type: "object",
+        description: "Extracted data from an invoice document.",
+        properties: {
+            vendor_name: {
+                type: "string",
+                description: "The name of the vendor or supplier.",
+            },
+            total_amount: {
+                type: "number",
+                description: "The total amount due on the invoice.",
+            },
+            invoice_date: {
+                type: "string",
+                description: "The date the invoice was issued (YYYY-MM-DD).",
+            },
+            line_items: {
+                type: "array",
+                description: "Individual items listed on the invoice.",
+                items: {
+                    type: "object",
+                    properties: {
+                        description: {
+                            type: "string",
+                            description: "Description of the item.",
+                        },
+                        quantity: { type: "number", description: "Quantity ordered." },
+                        unit_price: {
+                            type: "number",
+                            description: "Price per unit.",
+                        },
+                        total: {
+                            type: "number",
+                            description: "Total price for this line item.",
+                        },
+                    },
+                    required: ["description", "quantity", "unit_price", "total"],
+                    additionalProperties: false,
+                },
+            },
+        },
+        required: ["vendor_name", "total_amount", "invoice_date", "line_items"],
+        additionalProperties: false,
+    },
+    requestId: 42,
+};
+export const generateSchema = oc
+    .route({
+    summary: "Free: Generate a JSON Schema",
+    description: "Generate a JSON Schema from a natural language instruction. Use the resulting schema with the /extract endpoints to pull structured data from documents. Rate limited to 10 requests per IP per day.",
+    tags: ["Free"],
+    spec: (op) => {
+        op.security = [];
+        const reqBody = op.requestBody;
+        if (reqBody?.content) {
+            for (const mediaType of Object.values(reqBody.content)) {
+                mediaType.examples = requestExamples;
+            }
+        }
+        const responses = op.responses;
+        const okResponse = responses?.["200"];
+        if (okResponse?.content) {
+            for (const mediaType of Object.values(okResponse.content)) {
+                mediaType.example = responseExample;
+            }
+        }
+        return op;
+    },
+})
+    .input(inputSchema)
+    .output(outputSchema);

package/.tsc/lib/router/free/index.d.ts

@@ -0,0 +1,4 @@
+export { bankStatementParse } from "./bank-statement-parse";
+export { generateSchema } from "./generate-schema";
+export { publicationPdfUrl } from "./publication-pdf-url";
+export { rateLimitStatus } from "./rate-limit-status";

package/.tsc/lib/router/free/index.js

@@ -0,0 +1,4 @@
+export { bankStatementParse } from "./bank-statement-parse";
+export { generateSchema } from "./generate-schema";
+export { publicationPdfUrl } from "./publication-pdf-url";
+export { rateLimitStatus } from "./rate-limit-status";

package/.tsc/lib/router/free/publication-pdf-url.d.ts

@@ -0,0 +1,11 @@
+import { z } from "zod";
+export declare const publicationPdfUrl: import("@orpc/contract").ContractProcedureBuilderWithInputOutput<z.ZodObject<{
+    ids: z.ZodArray<z.ZodString>;
+}, z.core.$strip>, z.ZodObject<{
+    results: z.ZodArray<z.ZodObject<{
+        id: z.ZodString;
+        pdfUrl: z.ZodOptional<z.ZodString>;
+        title: z.ZodOptional<z.ZodString>;
+    }, z.core.$strip>>;
+    requestId: z.ZodNumber;
+}, z.core.$strip>, Record<never, never>, Record<never, never>>;

package/.tsc/lib/router/free/publication-pdf-url.js

@@ -0,0 +1,86 @@
+import { oc } from "@orpc/contract";
+import { z } from "zod";
+const inputSchema = z.object({
+    ids: z
+        .array(z.string().min(1, "Each ID must be a non-empty string"), {
+        message: "ids is required and must be an array of strings",
+    })
+        .min(1, "ids must contain at least 1 identifier")
+        .max(20, "ids must contain at most 20 identifiers")
+        .describe("Publication identifiers: DOI (e.g. 10.1371/journal.pone.0185809), PubMed ID (e.g. 33116299), ArXiv ID (e.g. 1706.03762), or other supported format"),
+});
+const resultSchema = z.object({
+    id: z.string().describe("The input publication identifier"),
+    pdfUrl: z
+        .string()
+        .optional()
+        .describe("Direct URL to the publication PDF, if available"),
+    title: z.string().optional().describe("Title of the publication"),
+});
+const outputSchema = z.object({
+    results: z.array(resultSchema).describe("Results for each input ID"),
+    requestId: z
+        .number()
+        .int()
+        .describe("Unique request identifier for this API call"),
+});
+const requestExamples = {
+    "Single ArXiv ID": {
+        summary: "Get PDF URL for an ArXiv paper",
+        value: {
+            ids: ["1706.03762"],
+        },
+    },
+    "Multiple IDs": {
+        summary: "Batch lookup with mixed identifier types",
+        value: {
+            ids: ["1706.03762", "33116299", "10.1371/journal.pone.0185809"],
+        },
+    },
+    "Single PubMed ID": {
+        summary: "Get PDF URL for a PubMed paper",
+        value: {
+            ids: ["33116299"],
+        },
+    },
+};
+const responseExample = {
+    results: [
+        {
+            id: "1706.03762",
+            pdfUrl: "https://arxiv.org/pdf/1706.03762v7",
+            title: "Attention Is All You Need",
+        },
+        {
+            id: "33116299",
+            pdfUrl: "https://www.nature.com/articles/s41586-020-2901-9.pdf",
+            title: "Identification of SARS-CoV-2 inhibitors using lung and colonic organoids",
+        },
+    ],
+    requestId: 42,
+};
+export const publicationPdfUrl = oc
+    .route({
+    summary: "Free: Get publication PDF URL",
+    description: "Look up publications by DOI, PubMed ID, ArXiv ID, or other identifiers and return their PDF URLs if available. Accepts up to 20 IDs per request. Rate limited to 20 requests per IP per day.",
+    tags: ["Free"],
+    spec: (op) => {
+        op.security = [];
+        const reqBody = op.requestBody;
+        if (reqBody?.content) {
+            for (const mediaType of Object.values(reqBody.content)) {
+                mediaType.examples = requestExamples;
+            }
+        }
+        const responses = op.responses;
+        const okResponse = responses?.["200"];
+        if (okResponse?.content) {
+            for (const mediaType of Object.values(okResponse.content)) {
+                mediaType.example = responseExample;
+            }
+        }
+        return op;
+    },
+})
+    .input(inputSchema)
+    .output(outputSchema);

package/.tsc/lib/router/free/rate-limit-status.d.ts

@@ -0,0 +1,18 @@
+import { z } from "zod";
+export declare const rateLimitStatus: import("@orpc/contract").ContractProcedureBuilderWithOutput<import("@orpc/contract").Schema<unknown, unknown>, z.ZodObject<{
+    bankStatementParse: z.ZodObject<{
+        limit: z.ZodNumber;
+        used: z.ZodNumber;
+        remaining: z.ZodNumber;
+    }, z.core.$strip>;
+    generateSchema: z.ZodObject<{
+        limit: z.ZodNumber;
+        used: z.ZodNumber;
+        remaining: z.ZodNumber;
+    }, z.core.$strip>;
+    publicationPdfUrl: z.ZodObject<{
+        limit: z.ZodNumber;
+        used: z.ZodNumber;
+        remaining: z.ZodNumber;
+    }, z.core.$strip>;
+}, z.core.$strip>, Record<never, never>, Record<never, never>>;

package/.tsc/lib/router/free/rate-limit-status.js

@@ -0,0 +1,36 @@
+import { oc } from "@orpc/contract";
+import { z } from "zod";
+const limitSchema = z.object({
+    limit: z.number().int().describe("Maximum requests allowed per day"),
+    used: z.number().int().describe("Requests used today"),
+    remaining: z.number().int().describe("Requests remaining today"),
+});
+const outputSchema = z.object({
+    bankStatementParse: limitSchema.describe("Bank statement parse limits"),
+    generateSchema: limitSchema.describe("JSON schema generator limits"),
+    publicationPdfUrl: limitSchema.describe("Publication PDF URL lookup limits"),
+});
+const responseExample = {
+    bankStatementParse: { limit: 5, used: 2, remaining: 3 },
+    generateSchema: { limit: 10, used: 0, remaining: 10 },
+    publicationPdfUrl: { limit: 20, used: 5, remaining: 15 },
+};
+export const rateLimitStatus = oc
+    .route({
+    method: "GET",
+    summary: "Free: Get rate limit status",
+    description: "Check the remaining rate limit for all free API endpoints. Limits are per IP and reset daily at midnight UTC.",
+    tags: ["Free"],
+    spec: (op) => {
+        op.security = [];
+        const responses = op.responses;
+        const okResponse = responses?.["200"];
+        if (okResponse?.content) {
+            for (const mediaType of Object.values(okResponse.content)) {
+                mediaType.example = responseExample;
+            }
+        }
+        return op;
+    },
+})
+    .output(outputSchema);

package/.tsc/lib/router/identity/ask.js

@@ -29,7 +29,7 @@ const askInputSchema = z.object({
         .describe("Base64-encoded identity document file content"),
     question: z
         .string()
-        .min(4)
+        .min(4, "question must be at least 4 characters")
         .describe("The question to answer about the identity document"),
     model: specializedModelSchema.describe("Model tier for answering the question. " +
         "'auto' (default): Automatically selects the best tier. " +

package/.tsc/lib/router/identity/extract.js

@@ -29,7 +29,7 @@ const extractInputSchema = z.object({
         .describe("Base64-encoded identity document file content"),
     prompt: z
         .string()
-        .min(4)
+        .min(4, "prompt must be at least 4 characters")
         .describe("The prompt instructing the AI how to extract data from the identity document"),
     schema: z
         .union([

package/.tsc/lib/router/identity/parse.js

@@ -2,7 +2,9 @@ import { oc } from "@orpc/contract";
 import { z } from "zod";
 import { getDefaultSpec } from "./get-default-spec";
 const specializedParseModelSchema = z
-    .enum(["pro", "max", "auto"])
+    .enum(["pro", "max", "auto"], {
+    message: "model must be one of: pro, max, auto",
+})
     .default("auto");
 const parseInputSchema = z.object({
     url: z

package/.tsc/lib/router/index.d.ts

@@ -1,6 +1,8 @@
+export * as academic from "./academic";
 export * as admin from "./admin";
 export * as authenticate from "./authenticate";
 export * as bankStatement from "./bankStatement";
 export * as document from "./document";
+export * as free from "./free";
 export * as identity from "./identity";
 export * as invoice from "./invoice";

package/.tsc/lib/router/index.js

@@ -1,6 +1,8 @@
+export * as academic from "./academic";
 export * as admin from "./admin";
 export * as authenticate from "./authenticate";
 export * as bankStatement from "./bankStatement";
 export * as document from "./document";
+export * as free from "./free";
 export * as identity from "./identity";
 export * as invoice from "./invoice";

package/.tsc/lib/router/invoice/ask.js

@@ -26,7 +26,7 @@ const askInputSchema = z.object({
     base64: z.string().optional().describe("Base64-encoded invoice file content"),
     question: z
         .string()
-        .min(4)
+        .min(4, "question must be at least 4 characters")
         .describe("The question to answer about the invoice"),
     model: specializedModelSchema.describe("Model tier for answering the question. " +
         "'auto' (default): Automatically selects the best tier. " +

package/.tsc/lib/router/invoice/extract.js

@@ -26,7 +26,7 @@ const extractInputSchema = z.object({
     base64: z.string().optional().describe("Base64-encoded invoice file content"),
     prompt: z
         .string()
-        .min(4)
+        .min(4, "prompt must be at least 4 characters")
         .describe("The prompt instructing the AI how to extract data from the invoice"),
     schema: z
         .union([

package/.tsc/lib/router/invoice/parse.js

@@ -2,7 +2,9 @@ import { oc } from "@orpc/contract";
 import { z } from "zod";
 import { getDefaultSpec } from "./get-default-spec";
 const specializedParseModelSchema = z
-    .enum(["pro", "max", "auto"])
+    .enum(["pro", "max", "auto"], {
+    message: "model must be one of: pro, max, auto",
+})
     .default("auto");
 const parseInputSchema = z.object({
     url: z

package/CHANGELOG.md

@@ -1,5 +1,19 @@
 # @pdfvector/instance-contract
 
+## 0.0.21
+### Patch Changes
+
+
+
+- [#125](https://github.com/phuctm97/pdfvector/pull/125) [`0eac9b7`](https://github.com/phuctm97/pdfvector/commit/0eac9b7afdbe9b8a41933fd1f92e48da51fb4480) Thanks [@khanhduyvt0101](https://github.com/khanhduyvt0101)! - Add free public APIs with IP rate limiting and interactive tool pages
+
+## 0.0.20
+### Patch Changes
+
+
+
+- [#120](https://github.com/phuctm97/pdfvector/pull/120) [`31b1d7b`](https://github.com/phuctm97/pdfvector/commit/31b1d7b22917ce8a57088132c05da7c4ec105b2b) Thanks [@khanhduyvt0101](https://github.com/khanhduyvt0101)! - Add academic APIs (search, fetch, findCitations) with proxy system, publication URLs, and playground improvements
+
 ## 0.0.19
 ### Patch Changes
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@pdfvector/instance-contract",
-  "version": "0.0.19",
+  "version": "0.0.21",
   "type": "module",
   "main": ".tsc/lib/index.js",
   "dependencies": {