document-mcp 1.2.2 → 1.3.0

package/README.md

@@ -52,7 +52,7 @@ Requires Node.js `>=20` and pnpm `11.6.0` (pinned via `packageManager` in `packa
 pnpm install
 pnpm build         # tsdown -> dist/ (ESM + CJS + .d.ts)
 pnpm typecheck     # tsc --noEmit
-pnpm lint          # eslint . --max-warnings 0
+pnpm lint          # eslint . --fix --cache --max-warnings 0
 pnpm test          # vitest run --project unit
 pnpm test:smoke    # rebuilds dist/, then spawns dist/bin.js as a real subprocess and drives it over genuine MCP stdio
 ```

package/dist/bin.js

@@ -8,7 +8,7 @@ import { readFontFace } from "pdf-codec";
 import { existsSync, readFileSync } from "node:fs";
 import { basename, join } from "node:path";
 //#region package.json
-var version = "1.2.2";
+var version = "1.3.0";
 //#endregion
 //#region src/io/document-input.ts
 const EXTENSION_TO_FORMAT = {
@@ -129,7 +129,8 @@ const ConvertDocumentInputSchema = z.object({
 	targetFormat: DocumentFormatSchema.describe("The format to convert the document to. Not every (source, targetFormat) pair is supported directly -- call list_document_conversions first to confirm this one is."),
 	output: DocumentOutputSchema.optional().describe("Where to write the converted document. Omit entirely to receive the bytes inline, base64-encoded, instead."),
 	fonts: z.array(FontInputSchema$1).optional().describe("Extra font faces to make available to the conversion, for a family the source document does not already embed. Only consulted by a conversion that runs a layout engine (a <format>-to-pdf conversion); every other conversion ignores this option entirely."),
-	onSubstitutionDiagnostics: z.boolean().optional().describe("When true, additionally report each individual font-substitution event as structured fontSubstitutions (which family/weight/style was requested, what it resolved to instead, and why). Every substitution is always reported as a plain diagnostic in `diagnostics` regardless of this flag -- this only controls whether the fuller, structured event is also collected.")
+	onSubstitutionDiagnostics: z.boolean().optional().describe("When true, additionally report each individual font-substitution event as structured fontSubstitutions (which family/weight/style was requested, what it resolved to instead, and why). Every substitution is always reported as a plain diagnostic in `diagnostics` regardless of this flag -- this only controls whether the fuller, structured event is also collected."),
+	images: z.record(z.string(), z.string()).optional().describe("A map from a markdown image destination (the part in the parentheses of ![](..)) to its base64-encoded PNG/JPEG bytes, for resolving a markdown source's own non-data: images. Only consulted by a markdown-sourced conversion; every other conversion ignores it. A destination absent from the map degrades to alt text, matching documents.js's own MarkdownImageResolver port -- an MCP caller has no filesystem context to read a relative path from, so any image that is not a data: URI must be supplied here explicitly to be embedded.")
 });
 const ConvertDocumentOutputSchema = z.object({
 	targetFormat: DocumentFormatSchema,
@@ -148,7 +149,7 @@ function registerConvertTools(server) {
 		description: "Converts a document from one supported format to another via documents.js's DocumentConverter port -- docx, pptx, xlsx, odt, odp, ods, odg, odf, markdown, and pdf. Not every (source, targetFormat) pair is supported directly (odf, for instance, only ever converts to pdf); call list_document_conversions first to see which pairs actually are.",
 		inputSchema: ConvertDocumentInputSchema,
 		outputSchema: ConvertDocumentOutputSchema
-	}, async ({ source, targetFormat, output, fonts, onSubstitutionDiagnostics }, ctx) => {
+	}, async ({ source, targetFormat, output, fonts, onSubstitutionDiagnostics, images }, ctx) => {
 		const { signal } = ctx.mcpReq;
 		const { bytes, format } = await resolveDocumentInput(source, { signal });
 		const fontSubstitutions = [];
@@ -166,7 +167,11 @@ function registerConvertTools(server) {
 				italic: font.italic,
 				bytes: base64ToBytes(font.bytesBase64)
 			})),
-			onFontSubstitution: onSubstitutionDiagnostics === true ? (substitution) => fontSubstitutions.push(substitution) : void 0
+			onFontSubstitution: onSubstitutionDiagnostics === true ? (substitution) => fontSubstitutions.push(substitution) : void 0,
+			images: images === void 0 ? void 0 : (destination) => {
+				const base64 = images[destination];
+				return base64 === void 0 ? void 0 : { bytes: base64ToBytes(base64) };
+			}
 		});
 		const resolvedOutput = await resolveDocumentOutput(result.document.bytes, output ?? {});
 		const structuredContent = {

package/dist/index.cjs

@@ -7,7 +7,7 @@ let pdf_codec = require("pdf-codec");
 let node_fs = require("node:fs");
 let node_path = require("node:path");
 //#region package.json
-var version = "1.2.2";
+var version = "1.3.0";
 //#endregion
 //#region src/io/document-input.ts
 const EXTENSION_TO_FORMAT = {
@@ -128,7 +128,8 @@ const ConvertDocumentInputSchema = zod.z.object({
 	targetFormat: documents_js.DocumentFormatSchema.describe("The format to convert the document to. Not every (source, targetFormat) pair is supported directly -- call list_document_conversions first to confirm this one is."),
 	output: DocumentOutputSchema.optional().describe("Where to write the converted document. Omit entirely to receive the bytes inline, base64-encoded, instead."),
 	fonts: zod.z.array(FontInputSchema$1).optional().describe("Extra font faces to make available to the conversion, for a family the source document does not already embed. Only consulted by a conversion that runs a layout engine (a <format>-to-pdf conversion); every other conversion ignores this option entirely."),
-	onSubstitutionDiagnostics: zod.z.boolean().optional().describe("When true, additionally report each individual font-substitution event as structured fontSubstitutions (which family/weight/style was requested, what it resolved to instead, and why). Every substitution is always reported as a plain diagnostic in `diagnostics` regardless of this flag -- this only controls whether the fuller, structured event is also collected.")
+	onSubstitutionDiagnostics: zod.z.boolean().optional().describe("When true, additionally report each individual font-substitution event as structured fontSubstitutions (which family/weight/style was requested, what it resolved to instead, and why). Every substitution is always reported as a plain diagnostic in `diagnostics` regardless of this flag -- this only controls whether the fuller, structured event is also collected."),
+	images: zod.z.record(zod.z.string(), zod.z.string()).optional().describe("A map from a markdown image destination (the part in the parentheses of ![](..)) to its base64-encoded PNG/JPEG bytes, for resolving a markdown source's own non-data: images. Only consulted by a markdown-sourced conversion; every other conversion ignores it. A destination absent from the map degrades to alt text, matching documents.js's own MarkdownImageResolver port -- an MCP caller has no filesystem context to read a relative path from, so any image that is not a data: URI must be supplied here explicitly to be embedded.")
 });
 const ConvertDocumentOutputSchema = zod.z.object({
 	targetFormat: documents_js.DocumentFormatSchema,
@@ -147,7 +148,7 @@ function registerConvertTools(server) {
 		description: "Converts a document from one supported format to another via documents.js's DocumentConverter port -- docx, pptx, xlsx, odt, odp, ods, odg, odf, markdown, and pdf. Not every (source, targetFormat) pair is supported directly (odf, for instance, only ever converts to pdf); call list_document_conversions first to see which pairs actually are.",
 		inputSchema: ConvertDocumentInputSchema,
 		outputSchema: ConvertDocumentOutputSchema
-	}, async ({ source, targetFormat, output, fonts, onSubstitutionDiagnostics }, ctx) => {
+	}, async ({ source, targetFormat, output, fonts, onSubstitutionDiagnostics, images }, ctx) => {
 		const { signal } = ctx.mcpReq;
 		const { bytes, format } = await resolveDocumentInput(source, { signal });
 		const fontSubstitutions = [];
@@ -165,7 +166,11 @@ function registerConvertTools(server) {
 				italic: font.italic,
 				bytes: (0, documents_js.base64ToBytes)(font.bytesBase64)
 			})),
-			onFontSubstitution: onSubstitutionDiagnostics === true ? (substitution) => fontSubstitutions.push(substitution) : void 0
+			onFontSubstitution: onSubstitutionDiagnostics === true ? (substitution) => fontSubstitutions.push(substitution) : void 0,
+			images: images === void 0 ? void 0 : (destination) => {
+				const base64 = images[destination];
+				return base64 === void 0 ? void 0 : { bytes: (0, documents_js.base64ToBytes)(base64) };
+			}
 		});
 		const resolvedOutput = await resolveDocumentOutput(result.document.bytes, output ?? {});
 		const structuredContent = {

package/dist/index.js

@@ -6,7 +6,7 @@ import { readFontFace } from "pdf-codec";
 import { existsSync, readFileSync } from "node:fs";
 import { basename, join } from "node:path";
 //#region package.json
-var version = "1.2.2";
+var version = "1.3.0";
 //#endregion
 //#region src/io/document-input.ts
 const EXTENSION_TO_FORMAT = {
@@ -127,7 +127,8 @@ const ConvertDocumentInputSchema = z.object({
 	targetFormat: DocumentFormatSchema.describe("The format to convert the document to. Not every (source, targetFormat) pair is supported directly -- call list_document_conversions first to confirm this one is."),
 	output: DocumentOutputSchema.optional().describe("Where to write the converted document. Omit entirely to receive the bytes inline, base64-encoded, instead."),
 	fonts: z.array(FontInputSchema$1).optional().describe("Extra font faces to make available to the conversion, for a family the source document does not already embed. Only consulted by a conversion that runs a layout engine (a <format>-to-pdf conversion); every other conversion ignores this option entirely."),
-	onSubstitutionDiagnostics: z.boolean().optional().describe("When true, additionally report each individual font-substitution event as structured fontSubstitutions (which family/weight/style was requested, what it resolved to instead, and why). Every substitution is always reported as a plain diagnostic in `diagnostics` regardless of this flag -- this only controls whether the fuller, structured event is also collected.")
+	onSubstitutionDiagnostics: z.boolean().optional().describe("When true, additionally report each individual font-substitution event as structured fontSubstitutions (which family/weight/style was requested, what it resolved to instead, and why). Every substitution is always reported as a plain diagnostic in `diagnostics` regardless of this flag -- this only controls whether the fuller, structured event is also collected."),
+	images: z.record(z.string(), z.string()).optional().describe("A map from a markdown image destination (the part in the parentheses of ![](..)) to its base64-encoded PNG/JPEG bytes, for resolving a markdown source's own non-data: images. Only consulted by a markdown-sourced conversion; every other conversion ignores it. A destination absent from the map degrades to alt text, matching documents.js's own MarkdownImageResolver port -- an MCP caller has no filesystem context to read a relative path from, so any image that is not a data: URI must be supplied here explicitly to be embedded.")
 });
 const ConvertDocumentOutputSchema = z.object({
 	targetFormat: DocumentFormatSchema,
@@ -146,7 +147,7 @@ function registerConvertTools(server) {
 		description: "Converts a document from one supported format to another via documents.js's DocumentConverter port -- docx, pptx, xlsx, odt, odp, ods, odg, odf, markdown, and pdf. Not every (source, targetFormat) pair is supported directly (odf, for instance, only ever converts to pdf); call list_document_conversions first to see which pairs actually are.",
 		inputSchema: ConvertDocumentInputSchema,
 		outputSchema: ConvertDocumentOutputSchema
-	}, async ({ source, targetFormat, output, fonts, onSubstitutionDiagnostics }, ctx) => {
+	}, async ({ source, targetFormat, output, fonts, onSubstitutionDiagnostics, images }, ctx) => {
 		const { signal } = ctx.mcpReq;
 		const { bytes, format } = await resolveDocumentInput(source, { signal });
 		const fontSubstitutions = [];
@@ -164,7 +165,11 @@ function registerConvertTools(server) {
 				italic: font.italic,
 				bytes: base64ToBytes(font.bytesBase64)
 			})),
-			onFontSubstitution: onSubstitutionDiagnostics === true ? (substitution) => fontSubstitutions.push(substitution) : void 0
+			onFontSubstitution: onSubstitutionDiagnostics === true ? (substitution) => fontSubstitutions.push(substitution) : void 0,
+			images: images === void 0 ? void 0 : (destination) => {
+				const base64 = images[destination];
+				return base64 === void 0 ? void 0 : { bytes: base64ToBytes(base64) };
+			}
 		});
 		const resolvedOutput = await resolveDocumentOutput(result.document.bytes, output ?? {});
 		const structuredContent = {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "document-mcp",
-  "version": "1.2.2",
+  "version": "1.3.0",
   "description": "MCP (Model Context Protocol) server exposing documents.js's document-conversion, .odb, metadata, and font tooling as MCP tools.",
   "type": "module",
   "repository": {
@@ -42,7 +42,7 @@
   "scripts": {
     "build": "tsdown",
     "prepublishOnly": "pnpm run lint && pnpm run typecheck && tsdown && publint && attw --pack",
-    "lint": "eslint . --max-warnings 0",
+    "lint": "eslint . --fix --cache --max-warnings 0",
     "typecheck": "tsc --noEmit",
     "test": "vitest run --project unit",
     "test:watch": "vitest --project unit",
@@ -69,7 +69,7 @@
   "packageManager": "pnpm@11.6.0",
   "dependencies": {
     "@modelcontextprotocol/server": "^2.0.0",
-    "documents.js": "^1.81.8",
+    "documents.js": "^1.90.0",
     "pdf-codec": "^1.11.0",
     "zod": "^4.2.0"
   },