AutoShorts is a local-first desktop application for turning long-form video or audio recordings into high-impact, vertical short-form clip candidates (9:16 portrait) with AI-powered viral moment ranking.
Rust
feat: add support for Gemini, OpenAI, and OpenRouter LLM providers
JayWebtech518964fb7ff2
diff --git a/src-tauri/src/lib.rs b/src-tauri/src/lib.rsindex bb2d37c..63308ea 100644--- a/src-tauri/src/lib.rs+++ b/src-tauri/src/lib.rs@@ -51,6 +51,9 @@ async fn environment_status(state: tauri::State<'_, AppState>) -> Result<Environhas_deepgram_key: std::env::var("DEEPGRAM_API_KEY").is_ok(),has_anthropic_key: std::env::var("ANTHROPIC_API_KEY").is_ok(),has_deepseek_key: std::env::var("DEEPSEEK_API_KEY").is_ok(),+ has_gemini_key: std::env::var("GEMINI_API_KEY").is_ok(),+ has_openai_key: std::env::var("OPENAI_API_KEY").is_ok(),+ has_openrouter_key: std::env::var("OPENROUTER_API_KEY").is_ok(),llm_provider,has_local_whisper_model,has_ollama,@@ -456,6 +459,30 @@ async fn generate_candidates(.await.map_err(to_command_error)?}+ "gemini" => {+ let key = api_key+ .or_else(|| std::env::var("GEMINI_API_KEY").ok())+ .ok_or_else(|| "Set GEMINI_API_KEY or supply Gemini API Key to generate candidates.".to_string())?;+ llm::detect_candidates_with_gemini(&normalized, &key)+ .await+ .map_err(to_command_error)?+ }+ "openai" => {+ let key = api_key+ .or_else(|| std::env::var("OPENAI_API_KEY").ok())+ .ok_or_else(|| "Set OPENAI_API_KEY or supply OpenAI API Key to generate candidates.".to_string())?;+ llm::detect_candidates_with_openai(&normalized, &key)+ .await+ .map_err(to_command_error)?+ }+ "openrouter" => {+ let key = api_key+ .or_else(|| std::env::var("OPENROUTER_API_KEY").ok())+ .ok_or_else(|| "Set OPENROUTER_API_KEY or supply OpenRouter API Key to generate candidates.".to_string())?;+ llm::detect_candidates_with_openrouter(&normalized, &key)+ .await+ .map_err(to_command_error)?+ }_ => {let key = api_key.or_else(|| std::env::var("DEEPSEEK_API_KEY").ok())diff --git a/src-tauri/src/llm.rs b/src-tauri/src/llm.rsindex 27907a3..0a2cd5f 100644--- a/src-tauri/src/llm.rs+++ b/src-tauri/src/llm.rs@@ -4,7 +4,6 @@ use serde_json::json;use crate::models::{CandidateDraft, NormalizedTranscript, TranscriptSegment};-#[derive(Debug, Deserialize)]struct AnthropicMessage {content: Vec<AnthropicContent>,@@ -86,6 +85,223 @@ Avoid rambling setup, context-dependent references, and pure filler. Return up tparse_candidate_json(&text, min_duration)}+#[derive(Debug, Deserialize)]+struct GeminiResponse {+ candidates: Vec<GeminiCandidate>,+}++#[derive(Debug, Deserialize)]+struct GeminiCandidate {+ content: GeminiContent,+}++#[derive(Debug, Deserialize)]+struct GeminiContent {+ parts: Vec<GeminiPart>,+}++#[derive(Debug, Deserialize)]+struct GeminiPart {+ text: Option<String>,+}++pub async fn detect_candidates_with_gemini(+ transcript: &NormalizedTranscript,+ api_key: &str,+) -> Result<Vec<CandidateDraft>> {+ let segments = compact_segments(&transcript.segments);+ let prompt = format!(+ "You are identifying the most viral moments and strongest short-form clip candidates from a long-form transcript. \+For each candidate, the clip must be self-contained, starting with an extremely engaging hook within the first 3 seconds (to capture immediate attention on social feeds), \+30-90 seconds long, and cut at clean sentence/thought boundaries. Favor highly shareable content: concrete stories, \+strong opinions, emotional turns, surprising or counter-intuitive claims, clear payoffs, and high-energy/dramatic peaks. \+Avoid rambling setup, context-dependent references, and pure filler. Return up to 10 candidates as JSON matching this schema: \+{{\"candidates\":[{{\"start\":0.0,\"end\":0.0,\"score\":0.0,\"hook\":\"...\",\"rationale\":\"...\"}}]}}\n\nTranscript:\n{segments}"+ );++ let model = std::env::var("GEMINI_MODEL").unwrap_or_else(|_| "gemini-2.5-flash".to_string());+ let url = format!(+ "https://generativelanguage.googleapis.com/v1beta/models/{}:generateContent?key={}",+ model, api_key+ );++ let response = reqwest::Client::new()+ .post(&url)+ .json(&json!({+ "contents": [+ {+ "parts": [+ {+ "text": prompt+ }+ ]+ }+ ],+ "generationConfig": {+ "responseMimeType": "application/json",+ "temperature": 0.2+ }+ }))+ .send()+ .await+ .context("calling Gemini")?;++ if !response.status().is_success() {+ let status = response.status();+ let body = response.text().await.unwrap_or_default();+ return Err(anyhow!("Gemini request failed ({status}): {body}"));+ }++ let res_body: GeminiResponse = response.json().await.context("parsing Gemini response")?;+ let text = res_body+ .candidates+ .first()+ .and_then(|c| c.content.parts.first())+ .and_then(|p| p.text.clone())+ .ok_or_else(|| anyhow!("Gemini response did not include content text"))?;++ let min_duration = if transcript.duration < 60.0 {+ (transcript.duration * 0.5).max(5.0)+ } else {+ 30.0+ };+ parse_candidate_json(&text, min_duration)+}++#[derive(Debug, Deserialize)]+struct ChatCompletionResponse {+ choices: Vec<ChatCompletionChoice>,+}++#[derive(Debug, Deserialize)]+struct ChatCompletionChoice {+ message: ChatCompletionMessage,+}++#[derive(Debug, Deserialize)]+struct ChatCompletionMessage {+ content: String,+}++pub async fn detect_candidates_with_openai(+ transcript: &NormalizedTranscript,+ api_key: &str,+) -> Result<Vec<CandidateDraft>> {+ let segments = compact_segments(&transcript.segments);+ let prompt = format!(+ "You are identifying the most viral moments and strongest short-form clip candidates from a long-form transcript. \+For each candidate, the clip must be self-contained, starting with an extremely engaging hook within the first 3 seconds (to capture immediate attention on social feeds), \+30-90 seconds long, and cut at clean sentence/thought boundaries. Favor highly shareable content: concrete stories, \+strong opinions, emotional turns, surprising or counter-intuitive claims, clear payoffs, and high-energy/dramatic peaks. \+Avoid rambling setup, context-dependent references, and pure filler. Return up to 10 candidates as JSON matching this schema: \+{{\"candidates\":[{{\"start\":0.0,\"end\":0.0,\"score\":0.0,\"hook\":\"...\",\"rationale\":\"...\"}}]}}\n\nTranscript:\n{segments}"+ );++ let model = std::env::var("OPENAI_MODEL").unwrap_or_else(|_| "gpt-4o-mini".to_string());++ let response = reqwest::Client::new()+ .post("https://api.openai.com/v1/chat/completions")+ .header("Authorization", format!("Bearer {api_key}"))+ .json(&json!({+ "model": model,+ "messages": [+ {+ "role": "user",+ "content": prompt,+ }+ ],+ "temperature": 0.2,+ "response_format": {+ "type": "json_object"+ }+ }))+ .send()+ .await+ .context("calling OpenAI")?;++ if !response.status().is_success() {+ let status = response.status();+ let body = response.text().await.unwrap_or_default();+ return Err(anyhow!("OpenAI request failed ({status}): {body}"));+ }++ let res_body: ChatCompletionResponse =+ response.json().await.context("parsing OpenAI response")?;+ let text = res_body+ .choices+ .first()+ .map(|c| c.message.content.clone())+ .ok_or_else(|| anyhow!("OpenAI response did not include choices content"))?;++ let min_duration = if transcript.duration < 60.0 {+ (transcript.duration * 0.5).max(5.0)+ } else {+ 30.0+ };+ parse_candidate_json(&text, min_duration)+}++pub async fn detect_candidates_with_openrouter(+ transcript: &NormalizedTranscript,+ api_key: &str,+) -> Result<Vec<CandidateDraft>> {+ let segments = compact_segments(&transcript.segments);+ let prompt = format!(+ "You are identifying the most viral moments and strongest short-form clip candidates from a long-form transcript. \+For each candidate, the clip must be self-contained, starting with an extremely engaging hook within the first 3 seconds (to capture immediate attention on social feeds), \+30-90 seconds long, and cut at clean sentence/thought boundaries. Favor highly shareable content: concrete stories, \+strong opinions, emotional turns, surprising or counter-intuitive claims, clear payoffs, and high-energy/dramatic peaks. \+Avoid rambling setup, context-dependent references, and pure filler. Return up to 10 candidates as JSON matching this schema: \+{{\"candidates\":[{{\"start\":0.0,\"end\":0.0,\"score\":0.0,\"hook\":\"...\",\"rationale\":\"...\"}}]}}\n\nTranscript:\n{segments}"+ );++ let model =+ std::env::var("OPENROUTER_MODEL").unwrap_or_else(|_| "google/gemini-2.5-flash".to_string());++ let response = reqwest::Client::new()+ .post("https://openrouter.ai/api/v1/chat/completions")+ .header("Authorization", format!("Bearer {api_key}"))+ .json(&json!({+ "model": model,+ "messages": [+ {+ "role": "user",+ "content": prompt,+ }+ ],+ "temperature": 0.2,+ "response_format": {+ "type": "json_object"+ }+ }))+ .send()+ .await+ .context("calling OpenRouter")?;++ if !response.status().is_success() {+ let status = response.status();+ let body = response.text().await.unwrap_or_default();+ return Err(anyhow!("OpenRouter request failed ({status}): {body}"));+ }++ let res_body: ChatCompletionResponse = response+ .json()+ .await+ .context("parsing OpenRouter response")?;+ let text = res_body+ .choices+ .first()+ .map(|c| c.message.content.clone())+ .ok_or_else(|| anyhow!("OpenRouter response did not include choices content"))?;++ let min_duration = if transcript.duration < 60.0 {+ (transcript.duration * 0.5).max(5.0)+ } else {+ 30.0+ };+ parse_candidate_json(&text, min_duration)+}+#[derive(Debug, Serialize)]struct ClaudeMessage<'a> {role: &'a str,@@ -164,7 +380,7 @@ pub async fn detect_candidates_with_local_llm(model_name: &str,) -> Result<Vec<CandidateDraft>> {let segments = compact_segments(&transcript.segments);-+let system_instructions = "You are identifying the most viral moments and strongest short-form clip candidates from a long-form transcript. \For each candidate, the clip must be self-contained, starting with an extremely engaging hook within the first 3 seconds (to capture immediate attention on social feeds), \30-90 seconds long, and cut at clean sentence/thought boundaries. \@@ -226,7 +442,10 @@ Ensure the 'start' and 'end' values correspond to actual timestamps in the transreturn Err(anyhow!("Local Ollama request failed ({status}): {body}"));}- let res_body: OllamaResponse = response.json().await.context("parsing local Ollama response")?;+ let res_body: OllamaResponse = response+ .json()+ .await+ .context("parsing local Ollama response")?;let min_duration = if transcript.duration < 60.0 {(transcript.duration * 0.5).max(5.0)} else {@@ -235,42 +454,6 @@ Ensure the 'start' and 'end' values correspond to actual timestamps in the transparse_candidate_json(&res_body.message.content, min_duration)}-pub fn demo_candidates(transcript: &NormalizedTranscript) -> Vec<CandidateDraft> {- transcript- .segments- .chunks(8)- .take(10)- .enumerate()- .filter_map(|(index, chunk)| {- let first = chunk.first()?;- let last = chunk.last()?;- let duration = last.end - first.start;- if duration < 12.0 {- return None;- }-- let end = if duration > 90.0 {- first.start + 90.0- } else {- last.end- };-- Some(CandidateDraft {- start: first.start,- end,- score: (0.86 - (index as f64 * 0.035)).max(0.55),- hook: first- .text- .split_whitespace()- .take(12)- .collect::<Vec<_>>()- .join(" "),- rationale: "Demo ranking generated without Claude. Use ANTHROPIC_API_KEY for production-quality moment detection.".to_string(),- })- })- .collect()-}-fn compact_segments(segments: &[TranscriptSegment]) -> String {segments.iter()@@ -294,12 +477,19 @@ fn parse_candidate_json(text: &str, min_duration: f64) -> Result<Vec<CandidateDr.trim();let val: serde_json::Value = serde_json::from_str(trimmed).context("parsing candidate JSON")?;-+let candidates_arr = if val.is_array() {val.as_array().cloned()} else if val.is_object() {let mut found_arr = None;- for key in &["candidates", "Candidates", "moments", "clips", "segments", "results"] {+ for key in &[+ "candidates",+ "Candidates",+ "moments",+ "clips",+ "segments",+ "results",+ ] {if let Some(arr) = val.get(*key).and_then(|v| v.as_array()) {found_arr = Some(arr.clone());break;@@ -390,12 +580,14 @@ fn parse_candidate_json(text: &str, min_duration: f64) -> Result<Vec<CandidateDrscore = 0.0;}- let hook = item.get("hook")+ let hook = item+ .get("hook").and_then(|v| v.as_str()).unwrap_or("").to_string();- let rationale = item.get("rationale")+ let rationale = item+ .get("rationale").and_then(|v| v.as_str()).unwrap_or("").to_string();@@ -413,8 +605,7 @@ fn parse_candidate_json(text: &str, min_duration: f64) -> Result<Vec<CandidateDr.clone().into_iter().filter(|candidate| {- (candidate.end - candidate.start) >= min_duration- && !candidate.hook.trim().is_empty()+ (candidate.end - candidate.start) >= min_duration && !candidate.hook.trim().is_empty()}).collect::<Vec<_>>();@@ -422,8 +613,7 @@ fn parse_candidate_json(text: &str, min_duration: f64) -> Result<Vec<CandidateDrcandidates = drafts.into_iter().filter(|candidate| {- (candidate.end - candidate.start) >= 5.0- && !candidate.hook.trim().is_empty()+ (candidate.end - candidate.start) >= 5.0 && !candidate.hook.trim().is_empty()}).collect::<Vec<_>>();}diff --git a/src-tauri/src/models.rs b/src-tauri/src/models.rsindex c6c1e46..87d8ddc 100644--- a/src-tauri/src/models.rs+++ b/src-tauri/src/models.rs@@ -9,6 +9,9 @@ pub struct EnvironmentStatus {pub has_deepgram_key: bool,pub has_anthropic_key: bool,pub has_deepseek_key: bool,+ pub has_gemini_key: bool,+ pub has_openai_key: bool,+ pub has_openrouter_key: bool,pub llm_provider: String,pub has_local_whisper_model: bool,pub has_ollama: bool,diff --git a/src/main.tsx b/src/main.tsxindex 4e25abe..13f561f 100644--- a/src/main.tsx+++ b/src/main.tsx@@ -32,6 +32,9 @@ type EnvironmentStatus = {hasDeepgramKey: boolean;hasAnthropicKey: boolean;hasDeepseekKey: boolean;+ hasGeminiKey: boolean;+ hasOpenaiKey: boolean;+ hasOpenrouterKey: boolean;llmProvider: string;hasLocalWhisperModel: boolean;hasOllama: boolean;@@ -125,8 +128,8 @@ function App() {const [transcriptionEngine, setTranscriptionEngine] = useState<"deepgram" | "local">(() => {return (localStorage.getItem("autoshorts_transcription_engine") as "deepgram" | "local") || "local";});- const [llmEngine, setLlmEngine] = useState<"claude" | "deepseek" | "local">(() => {- return (localStorage.getItem("autoshorts_llm_engine") as "claude" | "deepseek" | "local") || "local";+ const [llmEngine, setLlmEngine] = useState<"claude" | "deepseek" | "local" | "gemini" | "openai" | "openrouter">(() => {+ return (localStorage.getItem("autoshorts_llm_engine") as "claude" | "deepseek" | "local" | "gemini" | "openai" | "openrouter") || "local";});const [localLlmModel, setLocalLlmModel] = useState(() => {return localStorage.getItem("autoshorts_local_llm_model") || "llama3.2";@@ -140,6 +143,15 @@ function App() {const [deepseekKey, setDeepseekKey] = useState(() => {return localStorage.getItem("autoshorts_deepseek_key") || "";});+ const [geminiKey, setGeminiKey] = useState(() => {+ return localStorage.getItem("autoshorts_gemini_key") || "";+ });+ const [openaiKey, setOpenaiKey] = useState(() => {+ return localStorage.getItem("autoshorts_openai_key") || "";+ });+ const [openrouterKey, setOpenrouterKey] = useState(() => {+ return localStorage.getItem("autoshorts_openrouter_key") || "";+ });const [downloadingModelName, setDownloadingModelName] = useState<string | null>(null);const [modelDownloadStatus, setModelDownloadStatus] = useState("");@@ -170,6 +182,9 @@ function App() {const canUseCloudKey = environment?.hasDeepgramKey || deepgramKey.trim().length > 0;const canUseClaude = environment?.hasAnthropicKey || anthropicKey.trim().length > 0;const canUseDeepseek = environment?.hasDeepseekKey || deepseekKey.trim().length > 0;+ const canUseGemini = environment?.hasGeminiKey || geminiKey.trim().length > 0;+ const canUseOpenai = environment?.hasOpenaiKey || openaiKey.trim().length > 0;+ const canUseOpenrouter = environment?.hasOpenrouterKey || openrouterKey.trim().length > 0;const canTranscribe = transcriptionEngine === "local"? Boolean(environment?.hasLocalWhisperModel)@@ -177,7 +192,15 @@ function App() {const canUseActiveLlm = llmEngine === "local"? Boolean(environment?.hasOllama)- : (llmEngine === "claude" ? canUseClaude : canUseDeepseek);+ : llmEngine === "claude"+ ? canUseClaude+ : llmEngine === "deepseek"+ ? canUseDeepseek+ : llmEngine === "gemini"+ ? canUseGemini+ : llmEngine === "openai"+ ? canUseOpenai+ : canUseOpenrouter;useEffect(() => {void refresh();@@ -213,6 +236,18 @@ function App() {localStorage.setItem("autoshorts_deepseek_key", deepseekKey);}, [deepseekKey]);+ useEffect(() => {+ localStorage.setItem("autoshorts_gemini_key", geminiKey);+ }, [geminiKey]);++ useEffect(() => {+ localStorage.setItem("autoshorts_openai_key", openaiKey);+ }, [openaiKey]);++ useEffect(() => {+ localStorage.setItem("autoshorts_openrouter_key", openrouterKey);+ }, [openrouterKey]);+const pullModelDirectly = async (modelName: string) => {setDownloadingModelName(modelName);setModelDownloadProgress(0);@@ -311,7 +346,7 @@ function App() {async function runAutoPipeline(projectId: string) {setError(null);const env = await invoke<EnvironmentStatus>("environment_status");-+if (transcriptionEngine === "local") {if (!env.hasLocalWhisperModel) {setError("Import successful. Local Whisper GGML model (ggml-base.bin) is missing in your models directory. Please add it to start transcription.");@@ -331,12 +366,26 @@ function App() {return;}} else {- const activeKey = llmEngine === "claude" ? anthropicKey : deepseekKey;- const hasActiveKey = llmEngine === "claude"- ? (env.hasAnthropicKey || activeKey.trim().length > 0)- : (env.hasDeepseekKey || activeKey.trim().length > 0);+ const activeKey =+ llmEngine === "claude" ? anthropicKey :+ llmEngine === "deepseek" ? deepseekKey :+ llmEngine === "gemini" ? geminiKey :+ llmEngine === "openai" ? openaiKey :+ llmEngine === "openrouter" ? openrouterKey : "";+ const hasActiveKey =+ llmEngine === "claude" ? (env.hasAnthropicKey || activeKey.trim().length > 0) :+ llmEngine === "deepseek" ? (env.hasDeepseekKey || activeKey.trim().length > 0) :+ llmEngine === "gemini" ? (env.hasGeminiKey || activeKey.trim().length > 0) :+ llmEngine === "openai" ? (env.hasOpenaiKey || activeKey.trim().length > 0) :+ llmEngine === "openrouter" ? (env.hasOpenrouterKey || activeKey.trim().length > 0) : false;if (!hasActiveKey) {- setError(`Transcription complete. ${llmEngine === "claude" ? "Claude" : "DeepSeek"} API Key is missing. Please add it in settings to analyze viral moments.`);+ const engineName =+ llmEngine === "claude" ? "Claude" :+ llmEngine === "deepseek" ? "DeepSeek" :+ llmEngine === "gemini" ? "Gemini" :+ llmEngine === "openai" ? "OpenAI" :+ llmEngine === "openrouter" ? "OpenRouter" : "LLM";+ setError(`Transcription complete. ${engineName} API Key is missing. Please add it in settings to analyze viral moments.`);return;}}@@ -543,10 +592,10 @@ function App() {<div className="app-shell-container"><main className="app-shell"><aside className="sidebar">- <div- className="brand-row"- onClick={() => setDetail(null)}- style={{ cursor: "pointer" }}+ <div+ className="brand-row"+ onClick={() => setDetail(null)}+ style={{ cursor: "pointer" }}title="Go to Home Dashboard"><div className="brand-mark">@@ -596,7 +645,7 @@ function App() {<h2>{detail.project.name || fileName(detail.project.sourcePath)}</h2></div><div className="topbar-actions">- <button+ <buttonclassName={`icon-button settings-toggle ${showSettings ? "active" : ""}`}onClick={() => setShowSettings(!showSettings)}title="API Settings"@@ -622,16 +671,17 @@ function App() {<option value="local">Local Whisper (Offline)</option><option value="deepgram">Deepgram (Cloud)</option></select>- </label>- <label><span>LLM Engine</span><selectvalue={llmEngine}- onChange={(event) => setLlmEngine(event.target.value as "claude" | "deepseek" | "local")}+ onChange={(event) => setLlmEngine(event.target.value as any)}><option value="local">Ollama (Offline Local)</option><option value="claude">Claude (Cloud)</option><option value="deepseek">DeepSeek (Cloud)</option>+ <option value="gemini">Google Gemini (Cloud)</option>+ <option value="openai">OpenAI (Cloud)</option>+ <option value="openrouter">OpenRouter (Cloud)</option></select></label>{transcriptionEngine === "deepgram" && (@@ -667,6 +717,40 @@ function App() {/></label>)}+ {llmEngine === "gemini" && (+ <label>+ <span>Gemini API Key</span>+ <input+ value={geminiKey}+ onChange={(event) => setGeminiKey(event.target.value)}+ placeholder={environment?.hasGeminiKey ? "Loaded from env" : "Optional (Gemini API Key)"}+ type="password"+ />+ </label>+ )}+ {llmEngine === "openai" && (+ <label>+ <span>OpenAI API Key</span>+ <input+ value={openaiKey}+ onChange={(event) => setOpenaiKey(event.target.value)}+ placeholder={environment?.hasOpenaiKey ? "Loaded from env" : "Optional (OpenAI API Key)"}+ type="password"+ />+ </label>+ )}+ {llmEngine === "openrouter" && (+ <label>+ <span>OpenRouter API Key</span>+ <input+ value={openrouterKey}+ onChange={(event) => setOpenrouterKey(event.target.value)}+ placeholder={environment?.hasOpenrouterKey ? "Loaded from env" : "Optional (OpenRouter API Key)"}+ type="password"+ />+ </label>+ )}+{llmEngine === "local" && (<label><span>Ollama Model Name</span>@@ -677,9 +761,9 @@ function App() {placeholder="e.g. llama3.2, qwen2.5:7b"type="text"/>- <button- type="button"- className="icon-button"+ <button+ type="button"+ className="icon-button"style={{ minHeight: '36px', height: '36px' }}onClick={() => pullModelDirectly(localLlmModel)}>@@ -690,9 +774,9 @@ function App() {)}</div><div style={{ display: 'flex', justifyContent: 'flex-end', marginTop: '16px', borderTop: '1px solid var(--border-color)', paddingTop: '16px' }}>- <button- type="button"- className="icon-button"+ <button+ type="button"+ className="icon-button"style={{ background: 'rgba(239, 68, 68, 0.08)', borderColor: 'rgba(239, 68, 68, 0.2)', color: '#f87171' }}onClick={() => {if (window.confirm("Are you sure you want to reset your configuration and restart onboarding from scratch?")) {@@ -772,7 +856,12 @@ function App() {<div className="api-warning">{llmEngine === "local"? "⚠️ Ollama local server is not running at http://localhost:11434. Moment detection will not work."- : `⚠️ ${llmEngine === "claude" ? "Claude" : "DeepSeek"} API Key is missing. Viral moment identification will not work. Please add your key in API Settings.`}+ : `⚠️ ${llmEngine === "claude" ? "Claude" :+ llmEngine === "deepseek" ? "DeepSeek" :+ llmEngine === "gemini" ? "Gemini" :+ llmEngine === "openai" ? "OpenAI" :+ llmEngine === "openrouter" ? "OpenRouter" : "LLM"+ } API Key is missing. Viral moment identification will not work. Please add your key in API Settings.`}</div>)}@@ -814,7 +903,7 @@ function App() {{candidate.selected && <Check size={14} />}</div></div>-+<div className="candidate-body"><div className="candidate-meta"><span>{formatTime(candidate.startSec)} - {formatTime(candidate.endSec)}</span>@@ -822,7 +911,7 @@ function App() {</div><h4>{candidate.hook}</h4><p className="candidate-rationale">{candidate.rationale}</p>-+<div className="candidate-actions"><span className={`clip-status ${isCut ? "ready" : clip?.status === "error" ? "error" : ""}`}>{isCut ? "Cut ready" : clip?.status === "error" ? "Cut failed" : clip?.status ?? "Pending"}@@ -935,9 +1024,9 @@ function App() {<h3>Choose Caption Style</h3><p>Select how your automated captions should look on the portrait short-form video clips.</p></div>-+<div className="style-grid">- <div+ <divclassName={`style-card ${selectedStyle === "modern-box" ? "selected" : ""}`}onClick={() => setSelectedStyle("modern-box")}>@@ -948,7 +1037,7 @@ function App() {<div className="style-card-desc">Sleek white text inside a semi-transparent black background padding box. Highly readable.</div></div>- <div+ <divclassName={`style-card ${selectedStyle === "classic-outline" ? "selected" : ""}`}onClick={() => setSelectedStyle("classic-outline")}>@@ -959,7 +1048,7 @@ function App() {<div className="style-card-desc">Vibrant bold yellow text with a clean black outline. High-energy CapCut formatting.</div></div>- <div+ <divclassName={`style-card ${selectedStyle === "minimal-shadow" ? "selected" : ""}`}onClick={() => setSelectedStyle("minimal-shadow")}>@@ -970,7 +1059,7 @@ function App() {<div className="style-card-desc">Pure white text with a soft, elegant drop shadow. Unobtrusive and modern.</div></div>- <div+ <divclassName={`style-card ${selectedStyle === "vibrant-cyan" ? "selected" : ""}`}onClick={() => setSelectedStyle("vibrant-cyan")}>@@ -981,7 +1070,7 @@ function App() {<div className="style-card-desc">Vibrant tech cyan text with a black drop shadow for a clean look.</div></div>- <div+ <divclassName={`style-card ${selectedStyle === "vibrant-yellow-box" ? "selected" : ""}`}onClick={() => setSelectedStyle("vibrant-yellow-box")}>@@ -992,7 +1081,7 @@ function App() {<div className="style-card-desc">Bold black text inside a solid yellow padding box. Punchy and high visibility.</div></div>- <div+ <divclassName={`style-card ${selectedStyle === "vibrant-green" ? "selected" : ""}`}onClick={() => setSelectedStyle("vibrant-green")}>@@ -1003,7 +1092,7 @@ function App() {<div className="style-card-desc">High-energy neon green text with black borders and a drop shadow (Hormozi style).</div></div>- <div+ <divclassName={`style-card ${selectedStyle === "vibrant-red" ? "selected" : ""}`}onClick={() => setSelectedStyle("vibrant-red")}>@@ -1038,11 +1127,11 @@ function App() {<div className="download-loader"><Loader2 className="spin" size={48} /></div>-+<div className="progress-bar-container"><div className="progress-bar-fill" style={{ width: `${modelDownloadProgress}%` }}></div></div>-+<div className="download-stats"><span className="download-status">{modelDownloadStatus}</span><span className="download-percentage">{modelDownloadProgress}%</span>@@ -1123,11 +1212,11 @@ function Onboarding({}: OnboardingProps) {const [setupMode, setSetupMode] = useState<"choose" | "local" | "cloud" | "downloading">("choose");const [selectedModel, setSelectedModel] = useState<string>("llama3.2");-+const [dgKey, setDgKey] = useState(initialDeepgramKey);const [antKey, setAntKey] = useState(initialAnthropicKey);const [dsKey, setDsKey] = useState(initialDeepseekKey);-+const [downloadStatus, setDownloadStatus] = useState("Initializing download...");const [downloadProgress, setDownloadProgress] = useState(0);const [error, setError] = useState<string | null>(null);@@ -1150,12 +1239,12 @@ function Onboarding({setError("Please provide at least one LLM Key (Claude or DeepSeek).");return;}-+setTranscriptionEngine("deepgram");setDeepgramKey(dgKey.trim());localStorage.setItem("autoshorts_deepgram_key", dgKey.trim());localStorage.setItem("autoshorts_transcription_engine", "deepgram");-+if (antKey.trim()) {setLlmEngine("claude");setAnthropicKey(antKey.trim());@@ -1167,7 +1256,7 @@ function Onboarding({localStorage.setItem("autoshorts_deepseek_key", dsKey.trim());localStorage.setItem("autoshorts_llm_engine", "deepseek");}-+localStorage.setItem("autoshorts_onboarded", "true");onComplete();};@@ -1176,9 +1265,9 @@ function Onboarding({setError(null);setCheckingOllama(true);setDownloadProgress(0);-+await refreshEnv();-+let isOllamaRunning = false;try {const currentEnv = await invoke<EnvironmentStatus>("environment_status");@@ -1186,13 +1275,13 @@ function Onboarding({} catch (e) {// ignore}-+setCheckingOllama(false);if (!isOllamaRunning) {setSetupMode("downloading");setDownloadStatus("Ollama not found. Starting automatic installer...");-+try {const unlistenInstall = await listen<string>("ollama-install-status", (event) => {setDownloadStatus(event.payload);@@ -1225,18 +1314,18 @@ function Onboarding({});await invoke("pull_ollama_model", { modelName: selectedModel });-+unlisten();setTranscriptionEngine("local");setLlmEngine("local");setLocalLlmModel(selectedModel);-+localStorage.setItem("autoshorts_transcription_engine", "local");localStorage.setItem("autoshorts_llm_engine", "local");localStorage.setItem("autoshorts_local_llm_model", selectedModel);localStorage.setItem("autoshorts_onboarded", "true");-+onComplete();} catch (err) {setError(err instanceof Error ? err.message : String(err));@@ -1314,13 +1403,13 @@ function Onboarding({<div className="step-body"><h4>Set up local LLM (Ollama)</h4><p>- Ollama must be installed and running on your machine.+ Ollama must be installed and running on your machine.If you don't have it installed, you can download it from <a href="https://ollama.com" target="_blank" rel="noreferrer" style={{ color: 'var(--accent-primary)', textDecoration: 'underline' }}>ollama.com</a>.</p><p>Select a model to download:</p>-+<div className="model-cards">- <div+ <divclassName={`model-card ${selectedModel === "llama3.2" ? "active" : ""}`}onClick={() => setSelectedModel("llama3.2")}>@@ -1331,7 +1420,7 @@ function Onboarding({<p>Requires 8GB+ RAM. Recommended for standard setups. Fast and efficient.</p></div>- <div+ <divclassName={`model-card ${selectedModel === "qwen2.5:3b" ? "active" : ""}`}onClick={() => setSelectedModel("qwen2.5:3b")}>@@ -1342,7 +1431,7 @@ function Onboarding({<p>Requires 8GB+ RAM. Excellent coding and logical reasoning abilities.</p></div>- <div+ <divclassName={`model-card ${selectedModel === "qwen2.5:7b" ? "active" : ""}`}onClick={() => setSelectedModel("qwen2.5:7b")}>@@ -1359,9 +1448,9 @@ function Onboarding({<div className="onboarding-actions"><button type="button" className="icon-button" onClick={() => setSetupMode("choose")}>Back</button>- <button+ <buttontype="button"- className="primary-action compact"+ className="primary-action compact"onClick={startLocalSetup}disabled={checkingOllama}>@@ -1384,9 +1473,9 @@ function Onboarding({<div className="form-stack"><div className="input-group"><label>Deepgram API Key *</label>- <input- type="password"- value={dgKey}+ <input+ type="password"+ value={dgKey}onChange={(e) => setDgKey(e.target.value)}placeholder="Insert your Deepgram API Key (for transcription)"/>@@ -1394,9 +1483,9 @@ function Onboarding({<div className="input-group"><label>Claude API Key</label>- <input- type="password"- value={antKey}+ <input+ type="password"+ value={antKey}onChange={(e) => setAntKey(e.target.value)}placeholder="Insert your Anthropic API Key (moment detection)"/>@@ -1404,9 +1493,9 @@ function Onboarding({<div className="input-group"><label>DeepSeek API Key</label>- <input- type="password"- value={dsKey}+ <input+ type="password"+ value={dsKey}onChange={(e) => setDsKey(e.target.value)}placeholder="Insert your DeepSeek API Key (alternative moment detection)"/>@@ -1432,11 +1521,11 @@ function Onboarding({<div className="download-loader"><Loader2 className="spin" size={48} /></div>-+<div className="progress-bar-container"><div className="progress-bar-fill" style={{ width: `${downloadProgress}%` }}></div></div>-+<div className="download-stats"><span className="download-status">{downloadStatus}</span><span className="download-percentage">{downloadProgress}%</span>