feat: add SQLite database support as alternative to PostgreSQL

.env.example

@@ -3,18 +3,35 @@ SELF_HOSTED_MODE=true
 DISABLE_SIGNUP=true
 
 UPLOAD_PATH="./data/uploads"
-DATABASE_URL="postgresql://user@localhost:5432/taxhacker"
+
+# Database Configuration
+# Option 1: PostgreSQL (recommended for production)
+DATABASE_URL="postgresql://user:password@localhost:5432/taxhacker"
+
+# Option 2: SQLite (simpler, single-file database)
+# DATABASE_URL="file:./data/taxhacker.db"
+
+# Optional: Explicitly set database provider (auto-detected from DATABASE_URL)
+# DATABASE_PROVIDER="postgresql"  # or "sqlite"
+
+# Set to true to skip database readiness check on startup
+# Useful when using external PostgreSQL or if the check fails due to URL parsing issues
+SKIP_DB_CHECK=false
 
 # You can put it here or the app will ask you to enter it
 OPENAI_MODEL_NAME="gpt-4o-mini"
 OPENAI_API_KEY=""  # "sk-..."
 
-GOOGLE_MODEL_NAME="gemini-2.5-flash"
+GOOGLE_MODEL_NAME="gemini-2.0-flash"
 GOOGLE_API_KEY=""
 
 MISTRAL_MODEL_NAME="mistral-medium-latest"
 MISTRAL_API_KEY=""
 
+# Ollama (Local LLM) Configuration - requires a vision-capable model
+OLLAMA_MODEL_NAME="llava"
+OLLAMA_BASE_URL="http://localhost:11434"  # Ollama server URL
+
 # Auth Config
 BETTER_AUTH_SECRET="random-secret-key"  # please use any long random string here
 

.github/workflows/docker-latest.yml

@@ -43,5 +43,5 @@ jobs:
           platforms: linux/amd64,linux/arm64
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
-          cache-from: type=gha
-          cache-to: type=gha,mode=max
+          cache-from: type=gha,scope=node23
+          cache-to: type=gha,mode=max,scope=node23

.github/workflows/docker-release.yml

@@ -44,5 +44,5 @@ jobs:
           platforms: linux/amd64,linux/arm64
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
-          cache-from: type=gha
-          cache-to: type=gha,mode=max
+          cache-from: type=gha,scope=node23
+          cache-to: type=gha,mode=max,scope=node23

.nvmrc

@@ -0,0 +1 @@
+22

Dockerfile

@@ -3,6 +3,7 @@ FROM node:23-slim AS base
 # Default environment variables
 ENV PORT=7331
 ENV NODE_ENV=production
+ENV NEXT_TELEMETRY_DISABLED=1
 
 # Build stage
 FROM base AS builder
@@ -22,8 +23,8 @@ RUN npm ci
 # Copy source code
 COPY . .
 
-# Build the application
-RUN npm run build
+# Disable Next.js telemetry and build
+RUN npx next telemetry disable && npm run build
 
 # Production stage
 FROM base
@@ -47,6 +48,7 @@ RUN mkdir -p /app/upload
 COPY --from=builder /app/node_modules ./node_modules
 COPY --from=builder /app/package*.json ./
 COPY --from=builder /app/prisma ./prisma
+COPY --from=builder /app/scripts ./scripts
 COPY --from=builder /app/.next ./.next
 COPY --from=builder /app/public ./public
 COPY --from=builder /app/app ./app

ai/providers/llmProvider.ts

@@ -3,7 +3,7 @@ import { ChatGoogleGenerativeAI } from "@langchain/google-genai"
 import { ChatMistralAI } from "@langchain/mistralai"
 import { BaseMessage, HumanMessage } from "@langchain/core/messages"
 
-export type LLMProvider = "openai" | "google" | "mistral"
+export type LLMProvider = "openai" | "google" | "mistral" | "ollama"
 
 export interface LLMConfig {
   provider: LLMProvider
@@ -28,6 +28,44 @@ export interface LLMResponse {
   error?: string
 }
 
+// Known valid model names for each provider (for error messages)
+const VALID_MODELS: Record<LLMProvider, string[]> = {
+  openai: ["gpt-4o", "gpt-4o-mini", "gpt-4-turbo", "gpt-4", "gpt-3.5-turbo"],
+  google: ["gemini-2.0-flash", "gemini-2.0-flash-lite", "gemini-1.5-flash", "gemini-1.5-pro", "gemini-2.5-flash"],
+  mistral: ["mistral-large-latest", "mistral-medium-latest", "mistral-small-latest", "open-mistral-nemo"],
+  ollama: ["llava", "llava:13b", "llama3.2-vision", "bakllava", "moondream"],
+}
+
+function formatModelError(provider: LLMProvider, model: string, originalError: string): string {
+  const validModels = VALID_MODELS[provider]
+
+  // Check for common error patterns and provide helpful messages
+  if (
+    originalError.includes("does not support images") ||
+    originalError.includes("not found") ||
+    originalError.includes("model not found") ||
+    originalError.includes("Invalid model") ||
+    originalError.includes("404")
+  ) {
+    return `Model "${model}" for ${provider} is invalid or does not exist. Valid models include: ${validModels.join(", ")}. Please check your settings.`
+  }
+
+  if (originalError.includes("API key") || originalError.includes("authentication") || originalError.includes("401")) {
+    return `Invalid API key for ${provider}. Please verify your API key in settings.`
+  }
+
+  if (originalError.includes("rate limit") || originalError.includes("429")) {
+    return `Rate limit exceeded for ${provider}. Please try again later.`
+  }
+
+  if (originalError.includes("quota") || originalError.includes("insufficient")) {
+    return `Quota exceeded for ${provider}. Please check your billing/quota settings.`
+  }
+
+  // Return original error with model context
+  return `${provider} (model: ${model}) error: ${originalError}`
+}
+
 async function requestLLMUnified(config: LLMConfig, req: LLMRequest): Promise<LLMResponse> {
   try {
     const temperature = 0
@@ -50,6 +88,18 @@ async function requestLLMUnified(config: LLMConfig, req: LLMRequest): Promise<LL
         model: config.model,
         temperature: temperature,
       })
+    } else if (config.provider === "ollama") {
+      // Ollama uses OpenAI-compatible API
+      // config.apiKey contains the base URL (e.g., http://localhost:11434)
+      const baseUrl = config.apiKey.endsWith("/") ? config.apiKey.slice(0, -1) : config.apiKey
+      model = new ChatOpenAI({
+        configuration: {
+          baseURL: `${baseUrl}/v1`,
+        },
+        apiKey: "ollama", // Ollama doesn't require a real API key
+        model: config.model,
+        temperature: temperature,
+      })
     } else {
       return {
         output: {},
@@ -79,34 +129,53 @@ async function requestLLMUnified(config: LLMConfig, req: LLMRequest): Promise<LL
       provider: config.provider,
     }
   } catch (error: any) {
+    const originalError = error instanceof Error ? error.message : `${config.provider} request failed`
+    const formattedError = formatModelError(config.provider, config.model, originalError)
+
     return {
       output: {},
       provider: config.provider,
-      error: error instanceof Error ? error.message : `${config.provider} request failed`,
+      error: formattedError,
     }
   }
 }
 
 export async function requestLLM(settings: LLMSettings, req: LLMRequest): Promise<LLMResponse> {
+  const errors: string[] = []
+
   for (const config of settings.providers) {
     if (!config.apiKey || !config.model) {
-      console.info("Skipping provider:", config.provider)
+      console.info("Skipping provider:", config.provider, "(no API key or model configured)")
       continue
     }
-    console.info("Use provider:", config.provider)
+    console.info("Use provider:", config.provider, "with model:", config.model)
 
     const response = await requestLLMUnified(config, req)
 
     if (!response.error) {
       return response
     } else {
       console.error(response.error)
+      errors.push(response.error)
+    }
+  }
+
+  // Build a helpful error message
+  const configuredProviders = settings.providers.filter(p => p.apiKey && p.model)
+
+  if (configuredProviders.length === 0) {
+    return {
+      output: {},
+      provider: settings.providers[0]?.provider || "openai",
+      error: "No LLM providers configured. Please add an API key and model name in Settings > LLM Providers.",
     }
   }
 
+  // Include specific errors for each failed provider
+  const errorDetails = errors.length > 0 ? ` Errors: ${errors.join(" | ")}` : ""
   return {
     output: {},
     provider: settings.providers[0]?.provider || "openai",
-    error: "All LLM providers failed or are not configured",
+    error: `All LLM providers failed.${errorDetails}`,
   }
 }

app/(app)/apps/invoices/actions.ts

@@ -29,14 +29,14 @@ export async function generateInvoicePDF(data: InvoiceFormData): Promise<Uint8Ar
 }
 
 export async function addNewTemplateAction(user: User, template: InvoiceTemplate) {
-  const appData = (await getAppData(user, "invoices")) as InvoiceAppData | null
+  const appData = await getAppData<InvoiceAppData>(user, "invoices")
   const updatedTemplates = [...(appData?.templates || []), template]
   const appDataResult = await setAppData(user, "invoices", { ...appData, templates: updatedTemplates })
   return { success: true, data: appDataResult }
 }
 
 export async function deleteTemplateAction(user: User, templateId: string) {
-  const appData = (await getAppData(user, "invoices")) as InvoiceAppData | null
+  const appData = await getAppData<InvoiceAppData>(user, "invoices")
   if (!appData) return { success: false, error: "No app data found" }
 
   const updatedTemplates = appData.templates.filter((t) => t.id !== templateId)

app/(app)/apps/invoices/page.tsx

@@ -14,7 +14,7 @@ export default async function InvoicesApp() {
   const user = await getCurrentUser()
   const settings = await getSettings(user.id)
   const currencies = await getCurrencies(user.id)
-  const appData = (await getAppData(user, "invoices")) as InvoiceAppData | null
+  const appData = await getAppData<InvoiceAppData>(user, "invoices")
 
   return (
     <div>

app/(app)/files/actions.ts

@@ -1,5 +1,14 @@
 "use server"
 
+// File-like interface for form data uploads (avoids runtime File check issues in Node.js)
+interface UploadedFile {
+  name: string
+  size: number
+  type: string
+  lastModified: number
+  arrayBuffer(): Promise<ArrayBuffer>
+}
+
 import { ActionState } from "@/lib/actions"
 import { getCurrentUser, isSubscriptionExpired } from "@/lib/auth"
 import {
@@ -18,13 +27,13 @@ import path from "path"
 
 export async function uploadFilesAction(formData: FormData): Promise<ActionState<null>> {
   const user = await getCurrentUser()
-  const files = formData.getAll("files") as File[]
+  const files = formData.getAll("files") as UploadedFile[]
 
   // Make sure upload dir exists
   const userUploadsDirectory = getUserUploadsDirectory(user)
 
   // Check limits
-  const totalFileSize = files.reduce((acc, file) => acc + file.size, 0)
+  const totalFileSize = files.reduce((acc, file) => acc + (file?.size || 0), 0)
   if (!isEnoughStorageToUploadFile(user, totalFileSize)) {
     return { success: false, error: `Insufficient storage to upload these files` }
   }
@@ -39,7 +48,7 @@ export async function uploadFilesAction(formData: FormData): Promise<ActionState
   // Process each file
   const uploadedFiles = await Promise.all(
     files.map(async (file) => {
-      if (!(file instanceof File)) {
+      if (!file || typeof file !== "object" || !("size" in file) || !("arrayBuffer" in file)) {
         return { success: false, error: "Invalid file" }
       }
 

app/(app)/import/csv/actions.tsx

@@ -1,5 +1,13 @@
 "use server"
 
+// File-like interface for form data uploads (avoids runtime File check issues in Node.js)
+interface UploadedFile {
+  name: string
+  size: number
+  type: string
+  arrayBuffer(): Promise<ArrayBuffer>
+}
+
 import { ActionState } from "@/lib/actions"
 import { getCurrentUser } from "@/lib/auth"
 import { EXPORT_AND_IMPORT_FIELD_MAP } from "@/models/export_and_import"
@@ -12,7 +20,7 @@ export async function parseCSVAction(
   _prevState: ActionState<string[][]> | null,
   formData: FormData
 ): Promise<ActionState<string[][]>> {
-  const file = formData.get("file") as File
+  const file = formData.get("file") as UploadedFile | null
   if (!file) {
     return { success: false, error: "No file uploaded" }
   }

app/(app)/settings/actions.ts

@@ -14,7 +14,7 @@ import { uploadStaticImage } from "@/lib/uploads"
 import { codeFromName, randomHexColor } from "@/lib/utils"
 import { createCategory, deleteCategory, updateCategory } from "@/models/categories"
 import { createCurrency, deleteCurrency, updateCurrency } from "@/models/currencies"
-import { createField, deleteField, updateField } from "@/models/fields"
+import { createField, deleteField, updateField, updateFieldOrders } from "@/models/fields"
 import { createProject, deleteProject, updateProject } from "@/models/projects"
 import { SettingsMap, updateSettings } from "@/models/settings"
 import { updateUser } from "@/models/users"
@@ -57,8 +57,8 @@ export async function saveProfileAction(
 
   // Upload avatar
   let avatarUrl = user.avatar
-  const avatarFile = formData.get("avatar") as File | null
-  if (avatarFile instanceof File && avatarFile.size > 0) {
+  const avatarFile = formData.get("avatar")
+  if (avatarFile && typeof avatarFile === "object" && "size" in avatarFile && (avatarFile as Blob).size > 0) {
     try {
       const uploadedAvatarPath = await uploadStaticImage(user, avatarFile, "avatar.webp", 500, 500)
       avatarUrl = `/files/static/${path.basename(uploadedAvatarPath)}`
@@ -69,8 +69,8 @@ export async function saveProfileAction(
 
   // Upload business logo
   let businessLogoUrl = user.businessLogo
-  const businessLogoFile = formData.get("businessLogo") as File | null
-  if (businessLogoFile instanceof File && businessLogoFile.size > 0) {
+  const businessLogoFile = formData.get("businessLogo")
+  if (businessLogoFile && typeof businessLogoFile === "object" && "size" in businessLogoFile && (businessLogoFile as Blob).size > 0) {
     try {
       const uploadedBusinessLogoPath = await uploadStaticImage(user, businessLogoFile, "businessLogo.png", 500, 500)
       businessLogoUrl = `/files/static/${path.basename(uploadedBusinessLogoPath)}`
@@ -195,6 +195,7 @@ export async function addCategoryAction(userId: string, data: Prisma.CategoryCre
       name: validatedForm.data.name,
       llm_prompt: validatedForm.data.llm_prompt,
       color: validatedForm.data.color || "",
+      parentCode: validatedForm.data.parentCode || null,
     })
     revalidatePath("/settings/categories")
 
@@ -221,6 +222,7 @@ export async function editCategoryAction(userId: string, code: string, data: Pri
     name: validatedForm.data.name,
     llm_prompt: validatedForm.data.llm_prompt,
     color: validatedForm.data.color || "",
+    parentCode: validatedForm.data.parentCode || null,
   })
   revalidatePath("/settings/categories")
 
@@ -288,3 +290,17 @@ export async function deleteFieldAction(userId: string, code: string) {
   revalidatePath("/settings/fields")
   return { success: true }
 }
+
+export async function reorderFieldsAction(
+  fieldOrders: { code: string; order: number }[]
+): Promise<ActionState<void>> {
+  try {
+    const user = await getCurrentUser()
+    await updateFieldOrders(user.id, fieldOrders)
+    revalidatePath("/settings/fields")
+    return { success: true }
+  } catch (error) {
+    console.error("Failed to reorder fields:", error)
+    return { success: false, error: "Failed to reorder fields" }
+  }
+}

app/(app)/settings/backups/actions.ts

@@ -1,5 +1,13 @@
 "use server"
 
+// File-like interface for form data uploads (avoids runtime File check issues in Node.js)
+interface UploadedFile {
+  name: string
+  size: number
+  type: string
+  arrayBuffer(): Promise<ArrayBuffer>
+}
+
 import { ActionState } from "@/lib/actions"
 import { getCurrentUser } from "@/lib/auth"
 import { prisma } from "@/lib/db"
@@ -23,7 +31,7 @@ export async function restoreBackupAction(
 ): Promise<ActionState<BackupRestoreResult>> {
   const user = await getCurrentUser()
   const userUploadsDirectory = getUserUploadsDirectory(user)
-  const file = formData.get("file") as File
+  const file = formData.get("file") as UploadedFile | null
 
   if (!file || file.size === 0) {
     return { success: false, error: "No file provided" }