@chrischall/gemini-mcp 2.4.2 → 2.5.0

package/dist/tools/interact.js

@@ -5,7 +5,7 @@ import { ChainedRequest404Error } from '../client.js';
 import { slugify, baseName, resolveImagePath, writeSidecar, lookupOutputDir, readOutputImageAsInline } from '../images.js';
 import { resolveImageInputs } from '../inputs.js';
 import { findInteractionImages, latestInteractionId } from '../sidecar.js';
-import { emit, reportShape, resolveAspectRatio, orientationSchema, ASPECT_RATIOS, IMAGE_SIZES, MODEL_CHOICE_GUIDE, resolveVideoInput, videoPathSchema, timeoutMsSchema, timeoutRiskHint, idempotencyKeySchema, asyncSchema, maxWaitMsSchema, withProgressHeartbeat, assertLocalInputsAvailable, imagesUrlSchema, imagesFileUrisSchema, imagesBase64Schema } from './shared.js';
+import { emit, reportShape, resolveAspectRatio, orientationSchema, ASPECT_RATIOS, IMAGE_SIZES, MODEL_CHOICE_GUIDE, resolveVideoInput, videoPathSchema, timeoutMsSchema, timeoutRiskHint, idempotencyKeySchema, asyncSchema, maxWaitMsSchema, withProgressHeartbeat, assertLocalInputsAvailable, imagesUrlSchema, imagesFileUrisSchema, imagesBase64Schema, hostedR2KeysParam, hostedUploadNote } from './shared.js';
 import { bytesToBase64 } from '../bytes.js';
 import { fingerprintRequest } from '../jobs.js';
 import { attachCost } from '../pricing.js';
@@ -89,6 +89,7 @@ export function registerInteractTools(server, client) {
             'do NOT start a new interaction or re-send the image for each tweak. `continue_last: true` chains from this ' +
             "session's most recent interaction without threading the id. " +
             recoveryDescription +
+            hostedUploadNote(client.mediaSink) +
             'Output is JPEG. ' +
             confirmNote('Local file inputs are confirmed first'),
         annotations: { readOnlyHint: false, destructiveHint: true, openWorldHint: true },
@@ -108,6 +109,7 @@ export function registerInteractTools(server, client) {
                 .describe(`Paths to reference input images. Local paths must be inside GEMINI_REFERENCE_DIR (else GEMINI_UPLOAD_DIR) when set. ${NEW_REFERENCES_ONLY}`),
             images_url: imagesUrlSchema(`Reference images. ${NEW_REFERENCES_ONLY} Given`),
             images_file_uris: imagesFileUrisSchema(`Reference images. ${NEW_REFERENCES_ONLY} Given`),
+            ...hostedR2KeysParam(client.mediaSink, `Reference images. ${NEW_REFERENCES_ONLY} Given`),
             images_base64: imagesBase64Schema(`Reference images. ${NEW_REFERENCES_ONLY} Given`),
             model: z
                 .string()
@@ -223,7 +225,7 @@ export function registerInteractTools(server, client) {
             thinking_level: args.thinking_level, previous_interaction_id: previousInteractionId,
             google_search: args.google_search, search_types: args.search_types,
             images, images_base64: args.images_base64, from_clipboard: args.from_clipboard,
-            images_url: args.images_url, images_file_uris: args.images_file_uris,
+            images_url: args.images_url, images_file_uris: args.images_file_uris, images_r2_keys: args.images_r2_keys,
             video_url: args.video_url, video_path: args.video_path,
         });
         return client.session.jobs.dispatch({ toolName: 'gemini_interact', fingerprint, idempotencyKey: args.idempotency_key, async: args.async, waitMs: args.max_wait_ms }, async () => {

package/dist/tools/music.js

@@ -2,7 +2,7 @@ import { z } from 'zod';
 import { slugify, baseName } from '../images.js';
 import { resolveImageInputs } from '../inputs.js';
 import { DEFAULT_MUSIC_MODEL } from '../models.js';
-import { emitMedia, timeoutMsSchema, idempotencyKeySchema, asyncSchema, maxWaitMsSchema, withProgressHeartbeat, assertLocalInputsAvailable, imagesUrlSchema, imagesFileUrisSchema, imagesBase64Schema } from './shared.js';
+import { emitMedia, timeoutMsSchema, idempotencyKeySchema, asyncSchema, maxWaitMsSchema, withProgressHeartbeat, assertLocalInputsAvailable, imagesUrlSchema, imagesFileUrisSchema, imagesBase64Schema, hostedR2KeysParam, hostedUploadNote } from './shared.js';
 import { fingerprintRequest } from '../jobs.js';
 import { attachCost } from '../pricing.js';
 import { confirmLocalInputs, confirmNote, confirmTokenParam } from './_confirm.js';
@@ -29,6 +29,7 @@ export function registerMusicTools(server, client) {
             'lyria-3-pro-preview (longer-form). Output is MP3, written to disk (or returned inline). Single-turn: a track ' +
             'cannot be refined by a follow-up call, so put the whole brief in the prompt. Runs long — give it a `max_wait_ms` ' +
             'budget (or `async: true` + gemini_get_result on a local install), or raise `timeout_ms`. Needs a funded account. ' +
+            hostedUploadNote(client.mediaSink) +
             confirmNote('Local file inputs are confirmed first'),
         annotations: { readOnlyHint: false, destructiveHint: true, openWorldHint: true },
         inputSchema: z.object({
@@ -37,6 +38,7 @@ export function registerMusicTools(server, client) {
             images: z.array(z.string().min(1)).optional().describe('Optional reference image path(s) to condition the music. Local paths must be inside GEMINI_REFERENCE_DIR (else GEMINI_UPLOAD_DIR) when set.'),
             images_url: imagesUrlSchema('Reference images'),
             images_file_uris: imagesFileUrisSchema('Reference images'),
+            ...hostedR2KeysParam(client.mediaSink, 'Reference images'),
             images_base64: imagesBase64Schema(),
             from_clipboard: z.boolean().optional().describe('Use the image currently on the macOS clipboard as a reference'),
             filename: z.string().optional().describe('Base filename for the output audio (extension stripped; default: slugified prompt)'),
@@ -62,7 +64,7 @@ export function registerMusicTools(server, client) {
         const fingerprint = fingerprintRequest('gemini_music_generate', {
             model, prompt: args.prompt,
             images: args.images, images_base64: args.images_base64, from_clipboard: args.from_clipboard,
-            images_url: args.images_url, images_file_uris: args.images_file_uris,
+            images_url: args.images_url, images_file_uris: args.images_file_uris, images_r2_keys: args.images_r2_keys,
         });
         return client.session.jobs.dispatch({ toolName: 'gemini_music_generate', fingerprint, idempotencyKey: args.idempotency_key, async: args.async, waitMs: args.max_wait_ms }, async (ctx) => {
             const { inputs, report } = await resolveImageInputs(args, client);

package/dist/tools/shared.js

@@ -144,6 +144,32 @@ export function imagesR2KeysSchema(label = 'Reference images') {
         "generation's media[].r2_key. The server reads its own bucket — no bytes in the conversation, no ~48h expiry. " +
         'Hosted connector only.');
 }
+/**
+ * `images_r2_keys` for a tool, but ONLY where a key can resolve.
+ *
+ * An `r2_key` names an object in the hosted connector's blob store; a disk
+ * install has no store, so there the param could never work and would be
+ * standing `tools/list` cost on every request (see tool-surface-budget). So it
+ * is registered from the sink's capability, like interact's recovery text.
+ *
+ * The cast is runtime-honest: off the hosted connector the key is absent from
+ * the schema, zod strips it, and `args.images_r2_keys` is `undefined` — which
+ * is exactly what the declared optional type allows.
+ */
+export function hostedR2KeysParam(sink, label) {
+    return ((sink ?? DEFAULT_SINK).persistsFiles ? {} : { images_r2_keys: imagesR2KeysSchema(label) });
+}
+/**
+ * The description sentence that sends a hosted caller to the upload route.
+ * Empty on disk (where `images` paths work and no upload tool exists), so it
+ * costs nothing there.
+ */
+export function hostedUploadNote(sink) {
+    return (sink ?? DEFAULT_SINK).persistsFiles
+        ? ''
+        : 'For a reference photo you hold locally: gemini_get_upload_url, curl PUT it, then pass its r2_key in images_r2_keys ' +
+            '(no bytes in the conversation; reusable across calls). ';
+}
 /** Saved-character names attached to a generation (hosted connector only). */
 export const charactersSchema = z
     .array(z.string().min(1))

package/dist/tools/video.js

@@ -3,7 +3,7 @@ import { McpToolError } from '@chrischall/mcp-utils';
 import { slugify, baseName } from '../images.js';
 import { resolveImageInputs } from '../inputs.js';
 import { DEFAULT_VIDEO_MODEL } from '../models.js';
-import { emitMedia, reportShape, resolveAspectRatio, orientationSchema, timeoutMsSchema, idempotencyKeySchema, asyncSchema, maxWaitMsSchema, withProgressHeartbeat, assertLocalInputsAvailable, imagesUrlSchema, imagesFileUrisSchema, imagesBase64Schema } from './shared.js';
+import { emitMedia, reportShape, resolveAspectRatio, orientationSchema, timeoutMsSchema, idempotencyKeySchema, asyncSchema, maxWaitMsSchema, withProgressHeartbeat, assertLocalInputsAvailable, imagesUrlSchema, imagesFileUrisSchema, imagesBase64Schema, hostedR2KeysParam, hostedUploadNote } from './shared.js';
 import { fingerprintRequest } from '../jobs.js';
 import { attachCost } from '../pricing.js';
 import { confirmLocalInputs, confirmNote, confirmTokenParam } from './_confirm.js';
@@ -35,6 +35,7 @@ export function registerVideoTools(server, client) {
             'add ~3-10s each, to ~40s total). Cost scales with `resolution` — draft at 360p, keep at 1080p/4k. ' +
             'Output is written to disk as MP4 (video has no inline MCP block). Video runs long — give it a `max_wait_ms` budget ' +
             '(or `async: true` + gemini_get_result on a local install), or raise `timeout_ms`. Needs a funded account. ' +
+            hostedUploadNote(client.mediaSink) +
             confirmNote('Local file inputs are confirmed first'),
         annotations: { readOnlyHint: false, destructiveHint: true, openWorldHint: true },
         inputSchema: z.object({
@@ -46,6 +47,7 @@ export function registerVideoTools(server, client) {
             images: z.array(z.string().min(1)).optional().describe('Reference image path(s) for image_to_video / reference_to_video. Local paths must be inside GEMINI_REFERENCE_DIR (else GEMINI_UPLOAD_DIR) when set.'),
             images_url: imagesUrlSchema('Reference stills'),
             images_file_uris: imagesFileUrisSchema('Reference stills'),
+            ...hostedR2KeysParam(client.mediaSink, 'Reference stills'),
             images_base64: imagesBase64Schema(),
             from_clipboard: z.boolean().optional().describe('Use the image currently on the macOS clipboard as a reference'),
             filename: z.string().optional().describe('Base filename for the output video (extension stripped; default: slugified prompt)'),
@@ -95,7 +97,7 @@ export function registerVideoTools(server, client) {
             // where `delivery` is not (same bytes, different transport).
             model, prompt: args.prompt, aspect_ratio: aspectRatio, resolution: args.resolution, task: args.task,
             images: args.images, images_base64: args.images_base64, from_clipboard: args.from_clipboard,
-            images_url: args.images_url, images_file_uris: args.images_file_uris,
+            images_url: args.images_url, images_file_uris: args.images_file_uris, images_r2_keys: args.images_r2_keys,
             previous_interaction_id: previousInteractionId,
         });
         return client.session.jobs.dispatch({ toolName: 'gemini_video_generate', fingerprint, idempotencyKey: args.idempotency_key, async: args.async, waitMs: args.max_wait_ms }, async (ctx) => {

package/dist/version.js

@@ -1,2 +1,2 @@
 /** Single source of the server version. release-please bumps the literal below. */
-export const VERSION = '2.4.2'; // x-release-please-version
+export const VERSION = '2.5.0'; // x-release-please-version

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@chrischall/gemini-mcp",
-  "version": "2.4.2",
+  "version": "2.5.0",
   "mcpName": "io.github.chrischall/gemini-mcp",
   "description": "Gemini image-generation MCP server for Claude — developed and maintained by AI (Claude Code)",
   "author": "Claude Code (AI) <https://www.anthropic.com/claude>",

package/server.json

@@ -6,12 +6,12 @@
     "url": "https://github.com/chrischall/gemini-mcp",
     "source": "github"
   },
-  "version": "2.4.2",
+  "version": "2.5.0",
   "packages": [
     {
       "registryType": "npm",
       "identifier": "@chrischall/gemini-mcp",
-      "version": "2.4.2",
+      "version": "2.5.0",
       "transport": {
         "type": "stdio"
       },