refactor: Provider settings panel improve

settings/ProviderDetailPane.cpp

@@ -42,6 +42,11 @@ ProviderDetailPane::ProviderDetailPane(QWidget *parent)
     spp.setColor(QPalette::WindowText, Utils::creatorColor(Utils::Theme::PanelTextColorMid));
     m_sourcePathLabel->setPalette(spp);
 
+    m_protocolLabel = new QLabel(this);
+    m_protocolLabel->setPalette(spp);
+    m_protocolLabel->setToolTip(
+        tr("The client API (protocol) this provider speaks. Fixed when the provider is created."));
+
     m_editBtn = new QPushButton(tr("Edit…"), this);
     m_editBtn->setDefault(true);
     m_openInEditorBtn = new QPushButton(tr("Open in editor"), this);
@@ -89,6 +94,7 @@ ProviderDetailPane::ProviderDetailPane(QWidget *parent)
     headerLeft->setContentsMargins(0, 0, 0, 0);
     headerLeft->setSpacing(2);
     headerLeft->addLayout(titleRow);
+    headerLeft->addWidget(m_protocolLabel);
     headerLeft->addWidget(m_sourcePathLabel);
 
     auto *headerRow = new QHBoxLayout;
@@ -107,8 +113,6 @@ ProviderDetailPane::ProviderDetailPane(QWidget *parent)
 
     auto *identitySection = new SectionBox(tr("Identity"), this);
     m_nameEdit = new QLineEdit(this);
-    m_typeEdit = new QLineEdit(this);
-    m_typeEdit->setReadOnly(true);
     m_descriptionEdit = new QPlainTextEdit(this);
     m_descriptionEdit->setMaximumHeight(60);
     m_descriptionEdit->setReadOnly(true);
@@ -118,9 +122,6 @@ ProviderDetailPane::ProviderDetailPane(QWidget *parent)
     identityGrid->setVerticalSpacing(4);
     FormBuilder(identityGrid)
         .row(tr("Name:"), m_nameEdit)
-        .row(tr("Client API:"), m_typeEdit,
-             tr("The client API this provider speaks. "
-                "Cannot be changed after creation."))
         .row(tr("Description:"), m_descriptionEdit);
     identitySection->bodyLayout()->addLayout(identityGrid);
 
@@ -151,8 +152,10 @@ ProviderDetailPane::ProviderDetailPane(QWidget *parent)
     m_revealKeyBtn = new QToolButton(this);
     m_revealKeyBtn->setText(QStringLiteral("👁"));
     m_revealKeyBtn->setCheckable(true);
-    m_revealKeyBtn->setToolTip(tr("Show / hide API key"));
+    m_revealKeyBtn->setToolTip(tr("Show / hide the API key (loads the stored key)"));
     connect(m_revealKeyBtn, &QToolButton::toggled, this, [this](bool on) {
+        if (on && m_apiKeyEdit->text().isEmpty() && m_currentHasStoredKey)
+            emit apiKeyRevealRequested();
         m_apiKeyEdit->setEchoMode(on ? QLineEdit::Normal : QLineEdit::Password);
     });
     m_apiKeySaveBtn = new QPushButton(tr("Save key"), this);
@@ -307,7 +310,7 @@ void ProviderDetailPane::populate(const Providers::ProviderInstance &inst, bool
         inst.description.isEmpty() ? tr("No description provided.") : inst.description);
 
     m_nameEdit->setText(inst.name);
-    m_typeEdit->setText(inst.clientApi);
+    m_protocolLabel->setText(tr("via %1").arg(inst.clientApi));
     m_descriptionEdit->setPlainText(inst.description);
     m_urlEdit->setText(inst.url);
 
@@ -363,7 +366,7 @@ void ProviderDetailPane::clear()
     m_sourcePathLabel->clear();
     m_descriptionLabel->clear();
     m_nameEdit->clear();
-    m_typeEdit->clear();
+    m_protocolLabel->clear();
     m_descriptionEdit->clear();
     m_urlEdit->clear();
     m_apiKeyEdit->clear();
@@ -476,6 +479,16 @@ void ProviderDetailPane::changeEvent(QEvent *event)
     }
 }
 
+void ProviderDetailPane::showRevealedKey(const QString &key)
+{
+    if (key.isEmpty())
+        return;
+    m_apiKeyEdit->setText(key);
+    m_apiKeyEdit->setEchoMode(QLineEdit::Normal);
+    if (!m_revealKeyBtn->isChecked())
+        m_revealKeyBtn->setChecked(true);
+}
+
 void ProviderDetailPane::setEditing(bool on)
 {
     m_editing = on;

settings/ProviderDetailPane.hpp

@@ -32,6 +32,7 @@ public:
     void populate(const Providers::ProviderInstance &inst, bool hasStoredKey);
     void clear();
     void refreshKeyStatus(bool hasStoredKey);
+    void showRevealedKey(const QString &key);
     void setLaunchState(Providers::ProviderLauncher::State st, const QString &lastError);
     void resetLaunchTerminal(const QByteArray &scrollback);
     void appendLaunchBytes(const QByteArray &chunk);
@@ -44,6 +45,7 @@ signals:
     void deleteRequested();
     void apiKeySaveRequested(const QString &newKey);
     void apiKeyClearRequested();
+    void apiKeyRevealRequested();
     void launchStartRequested(const QString &providerName);
     void launchStopRequested(const QString &providerName);
     void launchRestartRequested(const QString &providerName);
@@ -74,7 +76,7 @@ private:
     QLabel *m_descriptionLabel = nullptr;
 
     QLineEdit *m_nameEdit = nullptr;
-    QLineEdit *m_typeEdit = nullptr;
+    QLabel *m_protocolLabel = nullptr;
     QPlainTextEdit *m_descriptionEdit = nullptr;
     QLineEdit *m_urlEdit = nullptr;
     QLabel *m_samplePreview = nullptr;

settings/ProvidersSettingsPage.cpp

@@ -125,6 +125,8 @@ public:
                 this, &ProvidersPageWidget::onApiKeySave);
         connect(m_detailPane, &ProviderDetailPane::apiKeyClearRequested,
                 this, &ProvidersPageWidget::onApiKeyClear);
+        connect(m_detailPane, &ProviderDetailPane::apiKeyRevealRequested,
+                this, &ProvidersPageWidget::onApiKeyReveal);
         connect(m_detailPane, &ProviderDetailPane::launchStartRequested,
                 this, &ProvidersPageWidget::onLaunchStart);
         connect(m_detailPane, &ProviderDetailPane::launchStopRequested,
@@ -447,6 +449,16 @@ private slots:
         m_detailPane->refreshKeyStatus(true);
     }
 
+    void onApiKeyReveal()
+    {
+        if (!m_factory || !m_secrets || m_currentName.isEmpty())
+            return;
+        const auto *inst = m_factory->instanceByName(m_currentName);
+        if (!inst || inst->apiKeyRef.isEmpty())
+            return;
+        m_detailPane->showRevealedKey(m_secrets->readKeySync(inst->apiKeyRef));
+    }
+
     void onApiKeyClear()
     {
         if (!m_factory || !m_secrets || m_currentName.isEmpty())

sources/providersConfig/claude.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "Claude"
 client_api  = "Claude"
-description = "Anthropic's hosted Claude API."
+description = "Cloud (Anthropic). Claude Opus/Sonnet/Haiku via the Messages API."
 
 url         = "https://api.anthropic.com"
 api_key_ref = "qodeassist/providers/Claude"

sources/providersConfig/codestral.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "Codestral"
 client_api  = "Codestral"
-description = "Mistral's Codestral FIM-capable code model API."
+description = "Cloud (Mistral). Codestral — a code model with native fill-in-the-middle, great for completion."
 
 url         = "https://codestral.mistral.ai"
 api_key_ref = "qodeassist/providers/Codestral"

sources/providersConfig/googleai.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "Google AI"
 client_api  = "Google AI"
-description = "Google AI Studio (Gemini) hosted API."
+description = "Cloud (Google). Gemini models via Google AI Studio."
 
 url         = "https://generativelanguage.googleapis.com/v1beta"
 api_key_ref = "qodeassist/providers/Google AI"

sources/providersConfig/llamacpp.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "llama.cpp"
 client_api  = "llama.cpp"
-description = "Local llama.cpp server (llama-server)."
+description = "Local (llama.cpp). Your own llama-server running GGUF models. Point the URL at your server (default :8080)."
 
 url         = "http://localhost:8080"
 api_key_ref = "qodeassist/providers/llama.cpp"

sources/providersConfig/lmstudio_chat.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "LM Studio (Chat Completions)"
 client_api  = "LM Studio (Chat Completions)"
-description = "Local LM Studio server over the /v1/chat/completions endpoint."
+description = "Local (LM Studio). The OpenAI Chat Completions API — the right pick for most LM Studio models."
 
 url         = "http://localhost:1234"
 api_key_ref = "qodeassist/providers/LM Studio (Chat Completions)"

sources/providersConfig/lmstudio_responses.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "LM Studio (Responses API)"
 client_api  = "LM Studio (Responses API)"
-description = "Local LM Studio server over the /v1/responses endpoint."
+description = "Local (LM Studio). The newer Responses API — use only if your model needs it, otherwise pick LM Studio (Chat Completions)."
 
 url         = "http://localhost:1234"
 api_key_ref = "qodeassist/providers/LM Studio (Responses API)"

sources/providersConfig/mistral.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "Mistral AI"
 client_api  = "Mistral AI"
-description = "Mistral's hosted chat / completions API."
+description = "Cloud (Mistral). Mistral chat/instruct models, incl. the Magistral reasoning model. For Mistral's FIM code model, use Codestral."
 
 url         = "https://api.mistral.ai"
 api_key_ref = "qodeassist/providers/Mistral AI"

sources/providersConfig/ollama_compat.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "Ollama (OpenAI-compatible)"
 client_api  = "Ollama (OpenAI-compatible)"
-description = "Local Ollama daemon spoken to via the OpenAI-compatible /v1 routes."
+description = "Local (Ollama). The OpenAI-compatible /v1 routes — for OpenAI-style agents. For everyday use prefer Ollama (Native)."
 
 url         = "http://localhost:11434"
 api_key_ref = "qodeassist/providers/Ollama (OpenAI-compatible)"

sources/providersConfig/ollama_native.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "Ollama (Native)"
 client_api  = "Ollama (Native)"
-description = "Default local Ollama daemon over its native /api/* endpoints."
+description = "Local (Ollama). The default — native /api endpoints with the best support for FIM, thinking and tools. Use this for the bundled Ollama agents."
 
 url         = "http://localhost:11434"
 api_key_ref = "qodeassist/providers/Ollama (Native)"

sources/providersConfig/openai_chat.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "OpenAI (Chat Completions)"
 client_api  = "OpenAI (Chat Completions)"
-description = "OpenAI's hosted /v1/chat/completions endpoint."
+description = "Cloud (OpenAI). GPT models via the standard Chat Completions API. The default OpenAI choice."
 
 url         = "https://api.openai.com/v1"
 api_key_ref = "qodeassist/providers/OpenAI (Chat Completions)"

sources/providersConfig/openai_compat.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "OpenAI Compatible"
 client_api  = "OpenAI Compatible"
-description = "Self-hosted OpenAI-compatible server (vLLM, TGI, ...). Edit the URL to match your deployment."
+description = "Self-hosted, OpenAI-compatible server (vLLM, TGI, LiteLLM, …). Point the URL at your deployment."
 
 url         = "http://localhost:1234/v1"
 api_key_ref = "qodeassist/providers/OpenAI Compatible"

sources/providersConfig/openai_responses.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "OpenAI (Responses API)"
 client_api  = "OpenAI (Responses API)"
-description = "OpenAI's hosted /v1/responses endpoint."
+description = "Cloud (OpenAI). GPT models via the newer Responses API (needed for some o-series reasoning models). If unsure, use OpenAI (Chat Completions)."
 
 url         = "https://api.openai.com/v1"
 api_key_ref = "qodeassist/providers/OpenAI (Responses API)"

sources/providersConfig/openrouter.toml

@@ -2,7 +2,7 @@ schema_version = 1
 
 name        = "OpenRouter"
 client_api  = "OpenRouter"
-description = "OpenRouter aggregator (https://openrouter.ai)."
+description = "Cloud aggregator (openrouter.ai) — a single gateway to models from many providers (Anthropic, OpenAI, Google, Meta, …)."
 
 url         = "https://openrouter.ai/api"
 api_key_ref = "qodeassist/providers/OpenRouter"