docs: add Zaps.ai to supported providers and documentation

README.md

@@ -39,8 +39,8 @@ cargo install moltis --git https://github.com/moltis-org/moltis
 
 ## Features
 
-- **Multi-provider LLM support** — OpenAI Codex, GitHub Copilot, and Local
-  LLM through a trait-based provider architecture
+- **Multi-provider LLM support** — Zaps.ai (Privacy Gateway), OpenAI Codex,
+  GitHub Copilot, and Local LLM through a trait-based provider architecture
 - **Streaming responses** — real-time token streaming for a responsive user
   experience, including when tools are enabled (tool calls stream argument
   deltas as they arrive)

crates/agents/src/providers/mod.rs

@@ -539,6 +539,13 @@ const OPENAI_COMPAT_PROVIDERS: &[OpenAiCompatDef] = &[
         default_base_url: "https://api.venice.ai/api/v1",
         models: &[],
     },
+    OpenAiCompatDef {
+        config_name: "zaps",
+        env_key: "ZAPS_API_KEY",
+        env_base_url_key: "ZAPS_BASE_URL",
+        default_base_url: "https://zaps.ai/v1",
+        models: &[],
+    },
     OpenAiCompatDef {
         config_name: "ollama",
         env_key: "OLLAMA_API_KEY",

crates/cli/src/doctor_commands.rs

@@ -123,6 +123,7 @@ const PROVIDER_ENV_MAP: &[(&str, &str, bool)] = &[
     ("minimax", "MINIMAX_API_KEY", false),
     ("moonshot", "MOONSHOT_API_KEY", false),
     ("venice", "VENICE_API_KEY", false),
+    ("zaps", "ZAPS_API_KEY", false),
     ("ollama", "OLLAMA_API_KEY", true),
     ("kimi-code", "KIMI_API_KEY", false),
 ];

crates/config/src/template.rs

@@ -88,8 +88,8 @@ offered = ["local-llm", "github-copilot", "openai", "anthropic", "ollama", "moon
 # All available providers:
 #   "anthropic", "openai", "gemini", "groq", "xai", "deepseek",
 #   "mistral", "openrouter", "cerebras", "minimax", "moonshot",
-#   "venice", "ollama", "local-llm", "openai-codex", "github-copilot",
-#   "kimi-code"
+#   "venice", "zaps", "ollama", "local-llm", "openai-codex",
+#   "github-copilot", "kimi-code"
 
 # ── Anthropic (Claude) ────────────────────────────────────────
 # [providers.anthropic]
@@ -145,6 +145,15 @@ models = ["gpt-5.3", "gpt-5.2"]              # Preferred models shown first
 # models = ["anthropic/claude-3.5-sonnet"]    # Any model IDs on OpenRouter
 # base_url = "https://openrouter.ai/api/v1"
 
+# ── Zaps.ai (PII redaction gateway) ────────────────────────────
+# Routes requests through Zaps.ai's privacy gateway which
+# automatically redacts PII before forwarding to upstream LLMs.
+# [providers.zaps]
+# enabled = true
+# api_key = "gk_..."                            # Or set ZAPS_API_KEY env var
+# models = ["gpt-4o", "deepseek-chat"]          # Any model supported by Zaps gateway
+# base_url = "https://zaps.ai/v1"
+
 # ── Moonshot (Kimi) ─────────────────────────────────────────
 [providers.moonshot]
 # enabled = true

crates/config/src/validate.rs

@@ -97,6 +97,7 @@ const KNOWN_PROVIDER_NAMES: &[&str] = &[
     "minimax",
     "moonshot",
     "venice",
+    "zaps",
     "ollama",
 ];
 

crates/gateway/src/provider_setup.rs

@@ -663,6 +663,15 @@ fn known_providers() -> Vec<KnownProvider> {
             requires_model: true,
             key_optional: false,
         },
+        KnownProvider {
+            name: "zaps",
+            display_name: "Zaps.ai",
+            auth_type: "api-key",
+            env_key: Some("ZAPS_API_KEY"),
+            default_base_url: Some("https://zaps.ai/v1"),
+            requires_model: true,
+            key_optional: false,
+        },
         KnownProvider {
             name: "ollama",
             display_name: "Ollama",

crates/gateway/src/server.rs

@@ -2150,6 +2150,7 @@ pub async fn start_gateway(
             ("minimax", "MINIMAX_API_KEY", "https://api.minimax.chat/v1"),
             ("moonshot", "MOONSHOT_API_KEY", "https://api.moonshot.ai/v1"),
             ("venice", "VENICE_API_KEY", "https://api.venice.ai/api/v1"),
+            ("zaps", "ZAPS_API_KEY", "https://zaps.ai/v1"),
         ];
 
         for (config_name, env_key, default_base) in EMBEDDING_CANDIDATES {

docs/src/providers.md

@@ -7,6 +7,7 @@ Configure providers through the web UI or directly in configuration files.
 
 | Provider | Auth | Notes |
 |----------|------|-------|
+| **Zaps.ai** | API Key | Privacy-first gateway (redacts PII) |
 | **OpenAI Codex** | OAuth | Codex-focused cloud models |
 | **GitHub Copilot** | OAuth | Requires active Copilot subscription |
 | **Local LLM** | Local runtime | Runs models on your machine |
@@ -54,6 +55,15 @@ model = "qwen2.5-coder-7b-q4_k_m"
 
 ## Provider Setup
 
+### Zaps.ai
+
+Zaps.ai acts as a privacy shield between you and upstream LLM providers, automatically redacting PII.
+
+1. Go to **Settings** -> **Providers** -> **Zaps.ai**.
+2. Enter your API Key (starts with `gk_`).
+3. (Optional) Set the Base URL if using a custom endpoint (default: `https://zaps.ai/v1`).
+4. Enter the model ID you wish to use (e.g., `deepseek-chat`).
+
 ### OpenAI Codex
 
 OpenAI Codex uses OAuth token import and OAuth-based access.