@goodandready/dsh-image-gen 0.8.6 → 0.8.8

package/lib/client.js

@@ -419,11 +419,11 @@ window.__ModuleLoader__.load({
     }
     const OUTPUT_FORMATS = ['png', 'jpeg', 'webp']
     const DELIVERY_MODES = ['link', 'image']
-    const PROVIDERS = ['fal', 'custom', 'codex', 'grok']
+    const PROVIDERS = ['fal', 'custom', 'codex', 'grok', 'local']
 
     /** Field specs in display order. */
     const FIELDS = [
-      { field: 'provider', spec: selectField('provider', PROVIDERS), kind: 'select', options: PROVIDERS, labelKey: 'f.provider', hintKey: 'f.providerHint' },
+      { field: 'provider', spec: selectField('provider', PROVIDERS), kind: 'select', options: PROVIDERS, labelKey: 'f.provider', hintKey: 'f.providerHint', providerNeeds: { fal: 'ключ FAL', custom: 'адрес + ключ', codex: 'подписка ChatGPT', grok: 'подписка Grok', local: 'адрес сервера' } },
       { field: 'model', spec: textField('model'), kind: 'text', labelKey: 'f.model', hintKey: 'f.modelHint', placeholderKey: 'p.model' },
       { field: 'apiKeyEnv', spec: textField('apiKeyEnv'), kind: 'text', labelKey: 'f.apiKeyEnv', hintKey: 'f.apiKeyEnvHint', placeholderKey: 'p.apiKeyEnv' },
       { field: 'baseURL', spec: textField('baseURL'), kind: 'text', labelKey: 'f.baseURL', hintKey: 'f.baseURLHint', placeholderKey: 'p.baseURL' },
@@ -457,7 +457,8 @@ window.__ModuleLoader__.load({
       }
       if (entry.kind === 'select') {
         const px = entry.sizePixels && entry.sizePixels[state.text] ? ` — ${entry.sizePixels[state.text]}` : ''
-        const hint = base.hint + px
+        const need = entry.providerNeeds && entry.providerNeeds[state.text] ? ` (нужно: ${entry.providerNeeds[state.text]})` : ''
+        const hint = base.hint + px + need
         return react_jsx_runtime.jsx(SelectField, { ...base, hint, options: entry.options, inheritLabel: t('f.inherit') })
       }
       return react_jsx_runtime.jsx(ValueField, { ...base, numeric: entry.kind === 'number', placeholder: entry.placeholderKey ? t(entry.placeholderKey) : '' })
@@ -742,6 +743,9 @@ window.__ModuleLoader__.load({
 
       const body = []
       if (prompt) body.push(react.createElement('div', { className: 'fal-prompt', key: 'p' }, prompt))
+      if (args.originalPrompt && args.originalPrompt !== prompt) {
+        body.push(react.createElement('div', { className: 'fal-meta', key: 'orig' }, `Исходный запрос: ${args.originalPrompt}`))
+      }
       if (failed) {
         body.push(react.createElement('div', { className: 'fal-err', key: 'e' }, parsed.text || 'неизвестная ошибка'))
       } else if (!parsed.attachment && parsed.url) {
@@ -772,6 +776,15 @@ window.__ModuleLoader__.load({
         body.push(react.createElement('div', { className: 'fal-meta', key: 'm' }, parsed.text))
       }
 
+      // Ссылка для скачивания: та же, что у картинки (наш роут или провайдерский url).
+      const downloadUrl = parsed.attachment
+        ? '/dsh-image-gen/image?id=' + encodeURIComponent(parsed.attachment.attachmentId ?? parsed.attachment.id ?? '')
+          + '&mt=' + encodeURIComponent(parsed.attachment.mediaType || 'image/png')
+          + '&b=' + encodeURIComponent(String(parsed.attachment.bytes ?? 0))
+          + '&w=' + encodeURIComponent(String(parsed.attachment.width ?? 0))
+          + '&h=' + encodeURIComponent(String(parsed.attachment.height ?? 0))
+        : parsed.url
+
       // Диагностика: метаданные генерации (prompt/seed/size/format), чтобы было
       // понятно, по каким параметрам получилась именно эта картинка.
       const metaParts = []
@@ -792,6 +805,14 @@ window.__ModuleLoader__.load({
           onClick: repeat,
         }, 'Повторить генерацию'))
       }
+      if (!running && !failed && downloadUrl) {
+        body.push(react.createElement('a', {
+          key: 'download',
+          className: 'fal-repeat',
+          href: downloadUrl,
+          download: 'generated-image.png',
+        }, 'Скачать'))
+      }
 
       return react.createElement('div', { className: 'fal-card' }, head, body)
     }

package/lib/index.js

@@ -15,7 +15,8 @@
 import z from '@deepseek-ai/schemastery'
 import { defineTool } from '@deepseek-ai/dsh-tools'
 import { credentialRef } from '@deepseek-ai/dsh-credentials'
-import { mkdir, writeFile } from 'node:fs/promises'
+import { mkdir, readFile, writeFile } from 'node:fs/promises'
+import os from 'node:os'
 import path from 'node:path'
 import {
   IMAGE_SIZES,
@@ -40,7 +41,7 @@ const NS = 'dsh-image-gen'
 // настройки лежат под старым именем. Оно читается как запасное и переносится
 // под новое имя один раз — молча, при первом запуске после обновления.
 const LEGACY_NS = 'dsh-fal-image-gen'
-export const inject = ['tools', 'attachments', 'credentials', 'webServer', 'settings']
+export const inject = ['tools', 'attachments', 'credentials', 'webServer', 'settings', 'llm']
 
 export const Config = z.object({
   provider: z
@@ -108,13 +109,86 @@ export const Config = z.object({
     .description('provider=custom: fixed size sent to the API, e.g. 1024x1024. '
       + 'Empty means the named size is translated automatically — set this only for an API picky about sizes.')
     .default(''),
+  localKind: z
+    .string()
+    .description('provider=local: which local API to use — "comfyui" or "a1111".')
+    .default('comfyui'),
+  localBaseURL: z
+    .string()
+    .description('provider=local: server address, e.g. http://127.0.0.1:8188 (ComfyUI) or http://127.0.0.1:7860 (A1111).')
+    .default(''),
+  localModel: z
+    .string()
+    .description('provider=local: model id (A1111) or workflow/schema name (ComfyUI). Empty means the server default.')
+    .default(''),
+  localSteps: z
+    .number()
+    .description('provider=local: sampling steps.')
+    .default(20),
+  localCfg: z
+    .number()
+    .description('provider=local: CFG scale.')
+    .default(7),
+  seedreamKeyEnv: z
+    .string()
+    .description('provider=seedream: credential reference / env var holding the API key.')
+    .default('SEEDREAM_API_KEY'),
+  seedreamModel: z
+    .string()
+    .description('provider=seedream: model id, e.g. seedream-4.0.')
+    .default('seedream-4.0'),
+  geminiKeyEnv: z
+    .string()
+    .description('provider=gemini: credential reference / env var holding the Google API key.')
+    .default('GEMINI_API_KEY'),
+  geminiModel: z
+    .string()
+    .description('provider=gemini: model id, e.g. gemini-2.0-flash-exp-image-generation.')
+    .default('gemini-2.0-flash-exp-image-generation'),
   outputDir: z
     .string()
     .description('Where generated images are saved. A relative path resolves against the session working directory; an absolute path is used as given.')
     .default('generated/images'),
+  historyLimit: z
+    .number()
+    .description('How many recent generations to keep in the in-memory history list.')
+    .default(50),
+  enhancePrompt: z
+    .boolean()
+    .description('Expand a short prompt into a detailed one through the chat model before generating. Off by default.')
+    .default(false),
+  enhanceModel: z
+    .string()
+    .description('Model used to enhance the prompt. Empty means the same model that leads the conversation.')
+    .default(''),
+  enhanceBelowChars: z
+    .number()
+    .description('Only enhance prompts shorter than this many characters.')
+    .default(200),
+  stylePreset: z
+    .string()
+    .description('Optional style suffix appended to the prompt before generation. Empty (default) means no style is applied and the prompt is used as-is.')
+    .default(''),
+  cacheBySeed: z
+    .boolean()
+    .description('If the same seed+prompt was already generated (present in history), return the cached result instead of generating again. Off by default.')
+    .default(false),
 })
 
 /** Keep a file stem safe for the filesystem. */
+/** Прочитать исходное изображение для правки: путь к файлу или attachment id. */
+export async function resolveSource(ctx, exec, ref) {
+  if (!ref) return undefined
+  const sessionCwd = exec.agent?.session?.header?.cwd
+  if (ref.startsWith('sha256:')) {
+    const stored = await ctx.attachments.readImage({ attachmentId: ref, mediaType: 'image/png', bytes: 0, width: 0, height: 0 })
+    return { bytes: Buffer.from(stored.data), mediaType: stored.ref?.mediaType || 'image/png' }
+  }
+  const p = path.resolve(sessionCwd || process.cwd(), ref)
+  const bytes = await readFile(p)
+  return { bytes, mediaType: 'image/png' }
+}
+
 export function slugify(input) {
   const stem = String(input ?? '')
     .toLowerCase()
@@ -165,6 +239,97 @@ function migrateLegacySettings(sctx, scope) {
   }
 }
 
+/** Каталог истории: общий для всех устройств, переживает рестарт. */
+export function historyFile() {
+  return path.join(process.env.DSH_HOME || path.join(os.homedir(), '.dsh'), 'dsh-image-gen', 'history.json')
+}
+
+/** Прочитать историю из файла (пусто, если файла нет). */
+/** Найти запись истории по seed+prompt, если файл ещё существует. */
+export async function findCached(entries, seed, prompt) {
+  if (seed === undefined) return undefined
+  return entries.find((e) => e.seed === seed && e.prompt === prompt)
+}
+
+/** Вернуть запись из кэша, если файл существует; иначе undefined. */
+export async function cachedResult(entry) {
+  if (!entry || !entry.path) return undefined
+  try { require('node:fs').accessSync(entry.path) } catch (e) { return undefined }
+  return {
+    path: entry.path,
+    url: entry.attachmentId ? `/dsh-image-gen/image?id=${encodeURIComponent(entry.attachmentId)}` : '',
+    width: 0,
+    height: 0,
+    seed: entry.seed,
+    prompt: entry.prompt,
+    format: 'png',
+    cached: true,
+    attachment: entry.attachmentId ? { attachmentId: entry.attachmentId, mediaType: 'image/png', bytes: 0, width: 0, height: 0, name: '' } : undefined,
+  }
+}
+
+export async function readHistory() {
+  try {
+    const raw = await readFile(historyFile(), 'utf-8')
+    const parsed = JSON.parse(raw)
+    return Array.isArray(parsed) ? parsed : []
+  } catch (e) {
+    return []
+  }
+}
+
+/** Записать историю в файл (перезапись целиком). */
+export async function writeHistory(entries) {
+  try {
+    await mkdir(path.dirname(historyFile()), { recursive: true })
+    await writeFile(historyFile(), JSON.stringify(entries, null, 2))
+  } catch (e) { /* история не критична */ }
+}
+
+/** Отфильтровать записи, чьи файлы ещё существуют; новые первыми. */
+export function filterHistory(entries, exists) {
+  return entries.filter((e) => exists(e.path)).slice(0, 50)
+}
+
+/** Собрать текст из итератора llm.stream (text-delta / block-end). */
+export async function collectText(iterable) {
+  let out = ''
+  let sawDelta = false
+  for await (const chunk of iterable) {
+    if (chunk && chunk.type === 'text-delta' && typeof chunk.text === 'string') {
+      out += chunk.text
+      sawDelta = true
+    } else if (
+      !sawDelta && chunk && chunk.type === 'block-end'
+      && chunk.block && chunk.block.type === 'text' && typeof chunk.block.text === 'string'
+    ) {
+      out += chunk.block.text
+    }
+  }
+  return out.trim()
+}
+
+/** Развернуть короткий промпт через чат-модель; при ошибке вернуть исходный. */
+export async function enhancePrompt(ctx, cfg, prompt, signal) {
+  if (!cfg.enhancePrompt) return { prompt, enhanced: false }
+  if (String(prompt).length >= (cfg.enhanceBelowChars || 200)) return { prompt, enhanced: false }
+  try {
+    const chunks = ctx.llm.stream({
+      ...(signal ? { signal } : {}),
+      ...(cfg.enhanceModel ? { model: cfg.enhanceModel } : {}),
+      messages: [
+        { role: 'system', content: 'You expand a short image-generation prompt into a detailed one. Reply with only the expanded prompt, no commentary.' },
+        { role: 'user', content: prompt },
+      ],
+    })
+    const text = await collectText(chunks)
+    if (!text) return { prompt, enhanced: false }
+    return { prompt: text, enhanced: true }
+  } catch (e) {
+    return { prompt, enhanced: false }
+  }
+}
+
 export function apply(ctx, config) {
 
   // The Web card edits this namespace; without registering it the card binds to
@@ -242,6 +407,27 @@ export function apply(ctx, config) {
     }), `dsh-image-gen: image route ${path}`)
   }
 
+  // История генераций: in-memory список, отфильтрованный по существованию файлов.
+  ctx.effect(() => ctx.webServer.register({
+    kind: 'exact',
+    path: '/dsh-image-gen/history',
+    handler: async (req, res) => {
+      if (req.method !== 'GET') {
+        res.writeHead(405, { 'Content-Type': 'application/json' })
+        res.end(JSON.stringify({ error: 'GET only' }))
+        return
+      }
+      const exists = (p) => { try { return require('node:fs').existsSync(p) } catch { return false } }
+      const entries = await readHistory()
+      const withThumbs = filterHistory(entries, exists).map((e) => ({
+        ...e,
+        thumbnailUrl: e.attachmentId ? `/dsh-image-gen/image?id=${encodeURIComponent(e.attachmentId)}` : '',
+      }))
+      res.writeHead(200, { 'Content-Type': 'application/json' })
+      res.end(JSON.stringify(withThumbs))
+    },
+  }), 'dsh-image-gen: history route')
+
   ctx.tools.register(
     defineTool({
       name: 'generate_image',
@@ -278,6 +464,11 @@ export function apply(ctx, config) {
           type: 'integer',
           description: 'Number of variations to generate in one call, 1-4 (default 1). Cost scales with count.',
         },
+        prompts: {
+          type: 'array',
+          items: { type: 'string' },
+          description: 'Optional list of prompts; one image is generated per prompt. When set, overrides count and the single prompt.',
+        },
         negative_prompt: {
           type: 'string',
           description: 'Optional text describing what NOT to draw; sent to providers that support it (ignored otherwise).',
@@ -286,6 +477,23 @@ export function apply(ctx, config) {
           type: 'number',
           description: 'Optional prompt adherence (e.g. 1-20); sent to providers that support it (ignored otherwise).',
         },
+        quality: {
+          type: 'string',
+          enum: ['low', 'medium', 'high'],
+          description: 'Optional quality for providers that support it (seedream, gemini). Ignored otherwise.',
+        },
+        source_image: {
+          type: 'string',
+          description: 'Optional path to an image file or an attachment id (sha256:...) to edit instead of drawing from scratch. Only providers that support image editing accept it; others refuse with a clear reason.',
+        },
+        mask: {
+          type: 'string',
+          description: 'Optional path or attachment id of a mask image (same size as source) to restrict the edit to a region. Only used when source_image is set and the provider supports masks.',
+        },
+        strength: {
+          type: 'number',
+          description: 'Optional edit strength (0-1) for providers that accept it; higher keeps more of the original.',
+        },
       },
       output: {
         schema: {
@@ -361,21 +569,26 @@ export function apply(ctx, config) {
         let subscriptionImages
         try { subscriptionImages = ctx.get && ctx.get('subscriptionImages') } catch (noService) { subscriptionImages = undefined }
 
+        const source = args.source_image ? await resolveSource(ctx, exec, args.source_image) : undefined
+        const mask = args.mask ? await resolveSource(ctx, exec, args.mask) : undefined
+        const enhanced = await enhancePrompt(ctx, cfg, args.prompt, exec.signal)
+        const styleSuffix = cfg.stylePreset ? `, ${cfg.stylePreset}` : ''
+        const effectivePrompt = enhanced.prompt + styleSuffix
         const providers = makeProviders(
           { fetchImpl: fetch, resolveKey: (ref) => resolveApiKey(ctx, ref), cfg, subscriptionImages },
-          { prompt: args.prompt, size, format, seed: args.seed, signal: exec.signal, negativePrompt: args.negative_prompt, guidanceScale: args.guidance_scale },
+          { prompt: effectivePrompt, size, format, seed: args.seed, signal: exec.signal, negativePrompt: args.negative_prompt, guidanceScale: args.guidance_scale, source, mask, strength: args.strength, quality: args.quality },
         )
         // Подписочные провайдеры (codex/grok) отдают {ok:false, reason} вместо исключения:
         // отказ должен дойти до модели текстом. Без проверки execute шёл дальше с пустыми
         // байтами, и пользователь получал битую карточку вместо внятного отказа.
         const guard = (generated) => { if (generated && generated.ok === false) throw new Error(generated.reason) }
-        const one = async (jobSeed, providerKey) => {
-          const gen = await providers[providerKey](jobSeed)
+        const one = async (jobSeed, providerKey, promptArg = effectivePrompt) => {
+          const gen = await providers[providerKey](jobSeed, promptArg)
           guard(gen)
           const bytes = gen.bytes
           const mediaType = gen.mediaType
           const extension = mediaType === 'image/jpeg' ? 'jpg' : mediaType === 'image/webp' ? 'webp' : 'png'
-          const stem = `${slugify(args.output_name || args.prompt)}-${Date.now().toString(36)}-${jobSeed}`
+          const stem = `${slugify(args.output_name || promptArg)}-${Date.now().toString(36)}-${jobSeed}`
           const name = `${stem}.${extension}`
 
           const attachment = await ctx.attachments.saveImage({
@@ -401,7 +614,7 @@ export function apply(ctx, config) {
           await writeFile(
             path.join(outDir, `${stem}.json`),
             JSON.stringify(buildSidecar({
-              prompt: args.prompt,
+              prompt: promptArg,
               size,
               format,
               seed: gen.seed,
@@ -412,16 +625,34 @@ export function apply(ctx, config) {
               mediaType,
               attachmentId: attachment.attachmentId,
               url: deliverAs === 'image' && gen.sourceUrl ? gen.sourceUrl : localUrl,
+              cost: gen.cost,
             }), null, 2),
           )
 
+          const entry = {
+            path: filePath,
+            prompt: promptArg,
+            provider,
+            size,
+            format,
+            seed: gen.seed,
+            createdAt: new Date().toISOString(),
+            attachmentId: attachment.attachmentId,
+          }
+          const current = await readHistory()
+          current.unshift(entry)
+          if (current.length > (cfg.historyLimit || 50)) current.length = cfg.historyLimit || 50
+          await writeHistory(current)
+
           return {
             path: filePath,
             url: deliverAs === 'image' && gen.sourceUrl ? gen.sourceUrl : localUrl,
             width: gen.width || attachment.width,
             height: gen.height || attachment.height,
             seed: gen.seed,
-            prompt: args.prompt,
+            prompt: promptArg,
+            originalPrompt: enhanced.enhanced ? args.prompt : undefined,
+            cost: gen.cost,
             format: mediaType.replace('image/', ''),
             attachment: {
               attachmentId: attachment.attachmentId,
@@ -434,15 +665,25 @@ export function apply(ctx, config) {
           }
         }
 
-        const count = normalizeCount(args.count)
         const seedBase = args.seed ?? Math.floor(Math.random() * 100000)
         // Fallback-цепочка: пробуем текущий провайдер, при отказе — следующий
         // по порядку (fal → custom → codex → grok), собирая причины отказов.
         const order = fallbackOrder(provider)
-        const generators = Object.fromEntries(PROVIDER_KEYS.map((k) => [k, (s) => one(s, k)]))
+        const generators = Object.fromEntries(PROVIDER_KEYS.map((k) => [k, (s, p) => one(s, k, p)]))
         const images = []
-        for (let i = 0; i < count; i += 1) {
-          images.push(await tryGenerate(generators, order, seedBase + i))
+        const historyEntries = cfg.cacheBySeed ? await readHistory() : []
+        const batchPrompts = Array.isArray(args.prompts) && args.prompts.length ? args.prompts : null
+        if (batchPrompts) {
+          for (let i = 0; i < batchPrompts.length; i += 1) {
+            const cached = cfg.cacheBySeed ? await cachedResult(findCached(historyEntries, seedBase + i, batchPrompts[i])) : undefined
+            images.push(cached || await tryGenerate(generators, order, seedBase + i, batchPrompts[i]))
+          }
+        } else {
+          const count = normalizeCount(args.count)
+          for (let i = 0; i < count; i += 1) {
+            const cached = cfg.cacheBySeed ? await cachedResult(findCached(historyEntries, seedBase + i, effectivePrompt)) : undefined
+            images.push(cached || await tryGenerate(generators, order, seedBase + i))
+          }
         }
         const first = images[0]
         return { ...first, images }

package/lib/providers.js

@@ -7,7 +7,7 @@
 // Сеть приходит параметром (fetchImpl), ключ — через resolveKey, поэтому оба
 // провайдера проверяются юнит-тестами без единого реального запроса.
 
-export const PROVIDER_KEYS = ['fal', 'custom', 'codex', 'grok']
+export const PROVIDER_KEYS = ['fal', 'custom', 'codex', 'grok', 'local', 'seedream', 'gemini']
 
 /** Привести count из аргумента инструмента к диапазону 1..4. */
 /** Порядок провайдеров для fallback: основной первым, остальные по PROVIDER_KEYS. */
@@ -21,11 +21,11 @@ export function fallbackOrder(primary) {
  * @param generators - массив функций (key, seed) => Promise<generated>.
  * @param order - порядок ключей, длина = generators.length.
  */
-export async function tryGenerate(generators, order, seed) {
+export async function tryGenerate(generators, order, seed, promptArg) {
   const refusals = []
   for (const key of order) {
     try {
-      const produced = await generators[key](seed)
+      const produced = await generators[key](seed, promptArg)
       return produced
     } catch (e) {
       refusals.push(`${key}: ${e.message || String(e)}`)
@@ -99,9 +99,9 @@ export function normalizeMediaType(contentType, fallbackFormat) {
 /** Метаданные одной генерации для sidecar-файла рядом с картинкой. */
 export function buildSidecar({
   prompt, size, format, seed, provider, deliverAs,
-  width, height, mediaType, attachmentId, url, createdAt = new Date().toISOString(),
+  width, height, mediaType, attachmentId, url, cost, createdAt = new Date().toISOString(),
 }) {
-  return { prompt, size, format, seed, provider, deliverAs, width, height, mediaType, attachmentId, url, createdAt }
+  return { prompt, size, format, seed, provider, deliverAs, width, height, mediaType, attachmentId, url, cost, createdAt }
 }
 
 /** Submit a generation job to the FAL queue. */
@@ -157,13 +157,33 @@ export async function pollStatus(fetchImpl, statusUrl, key, signal, pollInterval
  *   {bytes, mediaType, width, height, seed, sourceUrl}. Нулевые width/height
  *   означают «спроси у службы вложений»: она всё равно измеряет байты сама.
  */
+
+/** Именованный размер -> [width, height] для локальных API. */
+export function sizeToPixels(size) {
+  const px = SIZE_PIXELS[size]
+  if (!px) return [1024, 1024]
+  const [w, h] = px.split('x').map(Number)
+  return [w, h]
+}
+
+/** Собрать multipart-тело для /images/edits (OpenAI-совместимый edit). */
+export function buildEditForm({ source, mask, prompt, size, strength }) {
+  const form = new FormData()
+  form.append('image', new Blob([source.bytes], { type: source.mediaType || 'image/png' }), 'source.png')
+  if (mask) form.append('mask', new Blob([mask.bytes], { type: mask.mediaType || 'image/png' }), 'mask.png')
+  form.append('prompt', prompt)
+  if (size) form.append('size', size)
+  if (strength !== undefined) form.append('strength', String(strength))
+  return form
+}
+
 export function makeProviders(deps, job) {
   const { fetchImpl, resolveKey, cfg } = deps
-  const { prompt, size, format, seed, signal, negativePrompt, guidanceScale } = job
+  const { prompt, size, format, seed, signal, negativePrompt, guidanceScale, source, mask, strength, quality } = job
 
-  async function fal(seedArg = seed) {
+  async function fal(seedArg = seed, promptArg = prompt) {
     const key = await resolveKey(cfg.apiKeyEnv)
-    const body = { prompt, image_size: size, num_images: 1 }
+    const body = { prompt: promptArg, image_size: size, num_images: 1 }
     if (seedArg !== undefined) body.seed = seedArg
     if (negativePrompt !== undefined) body.negative_prompt = negativePrompt
     if (guidanceScale !== undefined) body.guidance_scale = guidanceScale
@@ -193,11 +213,12 @@ export function makeProviders(deps, job) {
       height: image.height ?? 0,
       seed: result.seed ?? seedArg ?? 0,
       sourceUrl: image.url,
+      cost: result.cost,
     }
   }
 
   // Любой OpenAI-совместимый API картинок. Один запрос вместо очереди FAL.
-  async function custom(seedArg = seed) {
+  async function custom(seedArg = seed, promptArg = prompt) {
     const base = String(cfg.customBaseURL || '').replace(/\/+$/, '')
     if (!base) throw new Error('Custom image provider: base URL is not configured (Settings → Image generation)')
     if (!cfg.customModel) throw new Error('Custom image provider: model is not configured')
@@ -210,17 +231,20 @@ export function makeProviders(deps, job) {
 
     // response_format не отправляем: новые модели OpenAI его отвергают, а ответ
     // всё равно приходит либо base64, либо ссылкой — принимаем оба.
-    const res = await fetchImpl(`${base}/images/generations`, {
+    const endpoint = source ? `${base}/images/edits` : `${base}/images/generations`
+    const res = await fetchImpl(endpoint, {
       method: 'POST',
       headers,
-      body: JSON.stringify({
-        model: cfg.customModel,
-        prompt,
-        n: 1,
-        size: cfg.customSize || SIZE_PIXELS[size] || size,
-        ...(negativePrompt !== undefined ? { negative_prompt: negativePrompt } : {}),
-        ...(guidanceScale !== undefined ? { guidance_scale: guidanceScale } : {}),
-      }),
+      body: source
+        ? buildEditForm({ source, mask, prompt: promptArg, size: cfg.customSize || SIZE_PIXELS[size] || size, strength })
+        : JSON.stringify({
+            model: cfg.customModel,
+            prompt: promptArg,
+            n: 1,
+            size: cfg.customSize || SIZE_PIXELS[size] || size,
+            ...(negativePrompt !== undefined ? { negative_prompt: negativePrompt } : {}),
+            ...(guidanceScale !== undefined ? { guidance_scale: guidanceScale } : {}),
+          }),
       signal,
     })
     const data = await res.json().catch(() => ({}))
@@ -267,7 +291,10 @@ export function makeProviders(deps, job) {
   // только запрос и разбор ответа. Токен сюда не попадает вовсе — служба
   // отдаёт готовую картинку, а не ключ доступа.
   function subscription(provider) {
-    return async function generate(seedArg = seed) {
+    return async function generate(seedArg = seed, promptArg = prompt) {
+      if (source) {
+        return { ok: false, provider, reason: `${provider}: не умеет править изображения — используйте fal, custom или local` }
+      }
       const images = deps.subscriptionImages
       if (!images || typeof images.generate !== 'function') {
         return {
@@ -280,7 +307,7 @@ export function makeProviders(deps, job) {
       try {
         produced = await images.generate({
           provider,
-          prompt,
+          prompt: promptArg,
           size: SUBSCRIPTION_SIZES[size] || '1024x1024',
           quality: cfg.subscriptionQuality || undefined,
           signal,
@@ -305,5 +332,158 @@ export function makeProviders(deps, job) {
     }
   }
 
-  return { fal, custom, codex: subscription('codex'), grok: subscription('grok') }
+  // Локальная генерация: ComfyUI (очередь + опрос) или Automatic1111 (txt2img).
+  async function local(seedArg = seed, promptArg = prompt) {
+    const base = String(cfg.localBaseURL || '').replace(/\/+$/, '')
+    if (!base) throw new Error('Local image provider: server address is not configured (Settings → Image generation)')
+    const [width, height] = sizeToPixels(size)
+    const kind = cfg.localKind === 'a1111' ? 'a1111' : 'comfyui'
+
+    if (kind === 'a1111') {
+      const body = {
+        prompt: promptArg,
+        negative_prompt: negativePrompt,
+        width,
+        height,
+        steps: cfg.localSteps ?? 20,
+        cfg_scale: cfg.localCfg ?? 7,
+        seed: seedArg ?? -1,
+      }
+      if (cfg.localModel) body.override_settings = { sd_model_checkpoint: cfg.localModel }
+      const res = await fetchImpl(`${base}/sdapi/v1/txt2img`, {
+        method: 'POST',
+        headers: { 'Content-Type': 'application/json' },
+        body: JSON.stringify(body),
+        signal,
+      })
+      if (!res.ok) {
+        throw new Error(`Local A1111 failed (HTTP ${res.status}): ${String(await res.text().catch(() => '')).slice(0, 300)}`)
+      }
+      const data = await res.json().catch(() => ({}))
+      const b64 = data.images && data.images[0]
+      if (!b64) throw new Error('Local A1111 returned no images')
+      return {
+        bytes: Buffer.from(b64, 'base64'),
+        mediaType: normalizeMediaType('image/png', format),
+        width,
+        height,
+        seed: seedArg ?? 0,
+        sourceUrl: '',
+      }
+    }
+
+    // ComfyUI: submit через /prompt, опрос /history/{prompt_id} до готовности.
+    const promptId = `dsh-${Date.now().toString(36)}-${Math.random().toString(36).slice(2, 8)}`
+    const workflow = {
+      prompt: {
+        '3': { class_type: 'KSampler', inputs: { seed: seedArg ?? 0, steps: cfg.localSteps ?? 20, cfg: cfg.localCfg ?? 7, sampler_name: 'euler', scheduler: 'normal', denoise: 1, model: ['4', 0], positive: ['6', 0], negative: ['7', 0], latent_image: ['5', 0] } },
+        '4': { class_type: 'CheckpointLoaderSimple', inputs: { ckpt_name: cfg.localModel || 'v1-5-pruned-emaonly.safetensors' } },
+        '5': { class_type: 'EmptyLatentImage', inputs: { width, height, batch_size: 1 } },
+        '6': { class_type: 'CLIPTextEncode', inputs: { text: promptArg, clip: ['4', 1] } },
+        '7': { class_type: 'CLIPTextEncode', inputs: { text: negativePrompt || '', clip: ['4', 1] } },
+        '8': { class_type: 'VAEDecode', inputs: { samples: ['3', 0], vae: ['4', 2] } },
+        '9': { class_type: 'SaveImage', inputs: { filename_prefix: 'dsh', images: ['8', 0] } },
+      },
+    }
+    const submit = await fetchImpl(`${base}/prompt`, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({ prompt: workflow, client_id: promptId }),
+      signal,
+    })
+    if (!submit.ok) {
+      throw new Error(`Local ComfyUI submit failed (HTTP ${submit.status}): ${String(await submit.text().catch(() => '')).slice(0, 300)}`)
+    }
+    const submitData = await submit.json().catch(() => ({}))
+    const pid = submitData.prompt_id
+    if (!pid) throw new Error('Local ComfyUI did not return a prompt_id')
+
+    const deadline = Date.now() + cfg.timeoutMs
+    for (;;) {
+      if (signal.aborted) throw new Error('Local ComfyUI generation cancelled')
+      if (Date.now() > deadline) throw new Error(`Local ComfyUI timed out after ${cfg.timeoutMs} ms`)
+      await new Promise((resolve) => {
+        const timer = setTimeout(resolve, cfg.pollIntervalMs || 2000)
+        signal.addEventListener('abort', () => { clearTimeout(timer); resolve() }, { once: true })
+      })
+      const hist = await fetchImpl(`${base}/history/${pid}`, { signal })
+      if (!hist.ok) continue
+      const histData = await hist.json().catch(() => ({}))
+      const entry = histData[pid]
+      if (entry && entry.outputs) {
+        const outputs = entry.outputs
+        const img = Object.values(outputs).flatMap((o) => o.images || []).find((i) => i && i.filename)
+        if (img) {
+          const dl = await fetchImpl(`${base}/view?filename=${encodeURIComponent(img.filename)}&subfolder=${encodeURIComponent(img.subfolder || '')}&type=${encodeURIComponent(img.type || 'output')}`, { signal })
+          if (!dl.ok) throw new Error(`Local ComfyUI download failed (HTTP ${dl.status})`)
+          return {
+            bytes: Buffer.from(await dl.arrayBuffer()),
+            mediaType: normalizeMediaType('image/png', format),
+            width,
+            height,
+            seed: seedArg ?? 0,
+            sourceUrl: '',
+          }
+        }
+      }
+    }
+  }
+
+  // Seedream (ByteDance): OpenAI-совместимый API картинок.
+  async function seedream(seedArg = seed, promptArg = prompt) {
+    const key = await resolveKey(cfg.seedreamKeyEnv)
+    const base = 'https://api.bytedanceapi.com/v1'
+    const res = await fetchImpl(`${base}/images/generations`, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json', Authorization: `Bearer ${key}` },
+      body: JSON.stringify({
+        model: cfg.seedreamModel || 'seedream-4.0',
+        prompt: promptArg,
+        n: 1,
+        size: SIZE_PIXELS[size] || size,
+        ...(negativePrompt !== undefined ? { negative_prompt: negativePrompt } : {}),
+        ...(quality !== undefined ? { quality } : {}),
+      }),
+      signal,
+    })
+    const data = await res.json().catch(() => ({}))
+    if (!res.ok) {
+      const detail = data?.error?.message || JSON.stringify(data).slice(0, 600)
+      throw new Error(`Seedream failed (HTTP ${res.status}): ${detail}`)
+    }
+    const item = data?.data?.[0]
+    if (!item) throw new Error('Seedream returned no images')
+    if (item.b64_json) {
+      return { bytes: Buffer.from(item.b64_json, 'base64'), mediaType: normalizeMediaType('image/png', format), width: 0, height: 0, seed: seedArg ?? 0, sourceUrl: '' }
+    }
+    if (!item.url) throw new Error('Seedream returned neither b64_json nor url')
+    const dl = await fetchImpl(item.url, { signal })
+    if (!dl.ok) throw new Error(`Seedream download failed (HTTP ${dl.status})`)
+    return { bytes: Buffer.from(await dl.arrayBuffer()), mediaType: normalizeMediaType('image/png', format), width: 0, height: 0, seed: seedArg ?? 0, sourceUrl: item.url }
+  }
+
+  // Gemini (Google): images.generate через GenAI API.
+  async function gemini(seedArg = seed, promptArg = prompt) {
+    const key = await resolveKey(cfg.geminiKeyEnv)
+    const model = cfg.geminiModel || 'gemini-2.0-flash-exp-image-generation'
+    const res = await fetchImpl(`https://generativelanguage.googleapis.com/v1beta/models/${model}:generateContent?key=${encodeURIComponent(key)}`, {
+      method: 'POST',
+      headers: { 'Content-Type': 'application/json' },
+      body: JSON.stringify({
+        contents: [{ parts: [{ text: promptArg }] }],
+        generationConfig: { responseModalities: ['IMAGE'], ...(quality !== undefined ? { imageConfig: { imageQuality: quality } } : {}) },
+      }),
+      signal,
+    })
+    const data = await res.json().catch(() => ({}))
+    if (!res.ok) {
+      const detail = data?.error?.message || JSON.stringify(data).slice(0, 600)
+      throw new Error(`Gemini failed (HTTP ${res.status}): ${detail}`)
+    }
+    const part = data?.candidates?.[0]?.content?.parts?.find((p) => p.inlineData?.data)
+    if (!part) throw new Error('Gemini returned no image')
+    return { bytes: Buffer.from(part.inlineData.data, 'base64'), mediaType: normalizeMediaType(part.inlineData.mimeType || 'image/png', format), width: 0, height: 0, seed: seedArg ?? 0, sourceUrl: '' }
+  }
+
+  return { fal, custom, codex: subscription('codex'), grok: subscription('grok'), local, seedream, gemini }
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@goodandready/dsh-image-gen",
-  "version": "0.8.6",
+  "version": "0.8.8",
   "description": "Image generation for DeepSeek Harness: a generate_image tool with pluggable providers — the FAL queue, any OpenAI-compatible images API, or a ChatGPT/Grok subscription with no API key at all. The picture is shown inline in the conversation; the model receives either a link (works with any chat model) or the image itself (needs dsh-vision-bridge or a vision-capable model).",
   "keywords": [
     "deepseek-harness",