diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index aa58055..d6f1ae1 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -7,6 +7,18 @@ on:
     branches: [main]
 
 jobs:
+  lint:
+    name: Lint
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6
+      - uses: actions/setup-node@v6
+        with:
+          node-version: 20
+          cache: npm
+      - run: npm ci
+      - run: npm run lint
+
   typecheck:
     name: Typecheck
     runs-on: ubuntu-latest
@@ -29,12 +41,12 @@ jobs:
           node-version: 20
           cache: npm
       - run: npm ci
-      - run: npm run test:run
+      - run: npx vitest run --coverage
 
   build:
     name: Build
     runs-on: ubuntu-latest
-    needs: [typecheck, test]
+    needs: [lint, typecheck, test]
     steps:
       - uses: actions/checkout@v6
       - uses: actions/setup-node@v6
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
index b39f252..782b416 100644
--- a/CONTRIBUTING.md
+++ b/CONTRIBUTING.md
@@ -42,6 +42,17 @@ npm run test:run      # Run tests once and exit
 
 All tests must pass before submitting a pull request.
 
+## Linting
+
+This project uses [Biome](https://biomejs.dev/) for linting and formatting:
+
+```bash
+npm run lint          # Check for lint and formatting issues
+npm run lint:fix      # Auto-fix issues
+```
+
+CI enforces linting — run `npm run lint` before pushing.
+
 ## Building and Running
 
 To build the TypeScript code:
@@ -58,6 +69,7 @@ npm start
 
 - Use TypeScript for all code
 - Follow ESM module conventions (this project uses `"type": "module"`)
+- Run `npm run lint` to check formatting and lint rules ([Biome](https://biomejs.dev/) enforces style automatically)
 - Keep code readable and maintainable
 - Comments should explain the "why", not the "what"
 
diff --git a/biome.json b/biome.json
new file mode 100644
index 0000000..1d442ca
--- /dev/null
+++ b/biome.json
@@ -0,0 +1,24 @@
+{
+  "$schema": "https://biomejs.dev/schemas/2.4.4/schema.json",
+  "files": {
+    "includes": ["src/**/*.ts"]
+  },
+  "linter": {
+    "enabled": true,
+    "rules": {
+      "recommended": true
+    }
+  },
+  "formatter": {
+    "enabled": true,
+    "indentStyle": "space",
+    "indentWidth": 2,
+    "lineWidth": 100
+  },
+  "javascript": {
+    "formatter": {
+      "quoteStyle": "double",
+      "trailingCommas": "all"
+    }
+  }
+}
diff --git a/package-lock.json b/package-lock.json
index 79d614b..afe6aad 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -1,12 +1,12 @@
 {
   "name": "tokencost-dev",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "tokencost-dev",
-      "version": "0.1.0",
+      "version": "0.1.1",
       "license": "MIT",
       "dependencies": {
         "@modelcontextprotocol/sdk": "^1.25.3",
@@ -17,6 +17,7 @@
         "tokencost-dev": "dist/index.js"
       },
       "devDependencies": {
+        "@biomejs/biome": "2.4.4",
         "@types/node": "^25.3.0",
         "@vitest/coverage-v8": "^4.0.18",
         "typescript": "^5.9.3",
@@ -86,6 +87,169 @@
         "node": ">=18"
       }
     },
+    "node_modules/@biomejs/biome": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/biome/-/biome-2.4.4.tgz",
+      "integrity": "sha512-tigwWS5KfJf0cABVd52NVaXyAVv4qpUXOWJ1rxFL8xF1RVoeS2q/LK+FHgYoKMclJCuRoCWAPy1IXaN9/mS61Q==",
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "bin": {
+        "biome": "bin/biome"
+      },
+      "engines": {
+        "node": ">=14.21.3"
+      },
+      "funding": {
+        "type": "opencollective",
+        "url": "https://opencollective.com/biome"
+      },
+      "optionalDependencies": {
+        "@biomejs/cli-darwin-arm64": "2.4.4",
+        "@biomejs/cli-darwin-x64": "2.4.4",
+        "@biomejs/cli-linux-arm64": "2.4.4",
+        "@biomejs/cli-linux-arm64-musl": "2.4.4",
+        "@biomejs/cli-linux-x64": "2.4.4",
+        "@biomejs/cli-linux-x64-musl": "2.4.4",
+        "@biomejs/cli-win32-arm64": "2.4.4",
+        "@biomejs/cli-win32-x64": "2.4.4"
+      }
+    },
+    "node_modules/@biomejs/cli-darwin-arm64": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-darwin-arm64/-/cli-darwin-arm64-2.4.4.tgz",
+      "integrity": "sha512-jZ+Xc6qvD6tTH5jM6eKX44dcbyNqJHssfl2nnwT6vma6B1sj7ZLTGIk6N5QwVBs5xGN52r3trk5fgd3sQ9We9A==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
+    "node_modules/@biomejs/cli-darwin-x64": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-darwin-x64/-/cli-darwin-x64-2.4.4.tgz",
+      "integrity": "sha512-Dh1a/+W+SUCXhEdL7TiX3ArPTFCQKJTI1mGncZNWfO+6suk+gYA4lNyJcBB+pwvF49uw0pEbUS49BgYOY4hzUg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "darwin"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
+    "node_modules/@biomejs/cli-linux-arm64": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-linux-arm64/-/cli-linux-arm64-2.4.4.tgz",
+      "integrity": "sha512-V/NFfbWhsUU6w+m5WYbBenlEAz8eYnSqRMDMAW3K+3v0tYVkNyZn8VU0XPxk/lOqNXLSCCrV7FmV/u3SjCBShg==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
+    "node_modules/@biomejs/cli-linux-arm64-musl": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-linux-arm64-musl/-/cli-linux-arm64-musl-2.4.4.tgz",
+      "integrity": "sha512-+sPAXq3bxmFwhVFJnSwkSF5Rw2ZAJMH3MF6C9IveAEOdSpgajPhoQhbbAK12SehN9j2QrHpk4J/cHsa/HqWaYQ==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
+    "node_modules/@biomejs/cli-linux-x64": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-linux-x64/-/cli-linux-x64-2.4.4.tgz",
+      "integrity": "sha512-R4+ZCDtG9kHArasyBO+UBD6jr/FcFCTH8QkNTOCu0pRJzCWyWC4EtZa2AmUZB5h3e0jD7bRV2KvrENcf8rndBg==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
+    "node_modules/@biomejs/cli-linux-x64-musl": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-linux-x64-musl/-/cli-linux-x64-musl-2.4.4.tgz",
+      "integrity": "sha512-gGvFTGpOIQDb5CQ2VC0n9Z2UEqlP46c4aNgHmAMytYieTGEcfqhfCFnhs6xjt0S3igE6q5GLuIXtdQt3Izok+g==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "linux"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
+    "node_modules/@biomejs/cli-win32-arm64": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-win32-arm64/-/cli-win32-arm64-2.4.4.tgz",
+      "integrity": "sha512-trzCqM7x+Gn832zZHgr28JoYagQNX4CZkUZhMUac2YxvvyDRLJDrb5m9IA7CaZLlX6lTQmADVfLEKP1et1Ma4Q==",
+      "cpu": [
+        "arm64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
+    "node_modules/@biomejs/cli-win32-x64": {
+      "version": "2.4.4",
+      "resolved": "https://registry.npmjs.org/@biomejs/cli-win32-x64/-/cli-win32-x64-2.4.4.tgz",
+      "integrity": "sha512-gnOHKVPFAAPrpoPt2t+Q6FZ7RPry/FDV3GcpU53P3PtLNnQjBmKyN2Vh/JtqXet+H4pme8CC76rScwdjDcT1/A==",
+      "cpu": [
+        "x64"
+      ],
+      "dev": true,
+      "license": "MIT OR Apache-2.0",
+      "optional": true,
+      "os": [
+        "win32"
+      ],
+      "engines": {
+        "node": ">=14.21.3"
+      }
+    },
     "node_modules/@esbuild/aix-ppc64": {
       "version": "0.27.3",
       "resolved": "https://registry.npmjs.org/@esbuild/aix-ppc64/-/aix-ppc64-0.27.3.tgz",
diff --git a/package.json b/package.json
index 34218f1..7b61dbc 100644
--- a/package.json
+++ b/package.json
@@ -19,6 +19,8 @@
     "dev": "tsc --watch",
     "start": "node dist/index.js",
     "prepublishOnly": "npm run build",
+    "lint": "biome check src/",
+    "lint:fix": "biome check --write src/",
     "test": "vitest",
     "test:run": "vitest run"
   },
@@ -47,6 +49,7 @@
     "zod": "^4.3.6"
   },
   "devDependencies": {
+    "@biomejs/biome": "2.4.4",
     "@types/node": "^25.3.0",
     "@vitest/coverage-v8": "^4.0.18",
     "typescript": "^5.9.3",
diff --git a/src/index.ts b/src/index.ts
index 55d838c..063ebe7 100644
--- a/src/index.ts
+++ b/src/index.ts
@@ -2,12 +2,9 @@
 
 import { Server } from "@modelcontextprotocol/sdk/server/index.js";
 import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
-import {
-  CallToolRequestSchema,
-  ListToolsRequestSchema,
-} from "@modelcontextprotocol/sdk/types.js";
-import { tools, executeTool } from "./tools.js";
+import { CallToolRequestSchema, ListToolsRequestSchema } from "@modelcontextprotocol/sdk/types.js";
 import { getModels } from "./pricing.js";
+import { executeTool, tools } from "./tools.js";
 
 const server = new Server(
   {
diff --git a/src/pricing.test.ts b/src/pricing.test.ts
index 01c8508..bd6593b 100644
--- a/src/pricing.test.ts
+++ b/src/pricing.test.ts
@@ -1,5 +1,5 @@
-import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
-import { existsSync, readFileSync, writeFileSync, mkdirSync } from "node:fs";
+import { existsSync, mkdirSync, writeFileSync } from "node:fs";
+import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
 
 // Mock fs so disk cache doesn't interfere
 vi.mock("node:fs", () => ({
@@ -303,9 +303,7 @@ describe("pricing module", () => {
       });
 
       const { refreshPrices } = await loadPricing();
-      await expect(refreshPrices()).rejects.toThrow(
-        "Failed to fetch pricing data: 503",
-      );
+      await expect(refreshPrices()).rejects.toThrow("Failed to fetch pricing data: 503");
     });
   });
 
@@ -435,10 +433,9 @@ describe("pricing module", () => {
       const { refreshPrices } = await loadPricing();
       await refreshPrices();
 
-      expect(mkdirSync).toHaveBeenCalledWith(
-        expect.stringContaining(".cache"),
-        { recursive: true },
-      );
+      expect(mkdirSync).toHaveBeenCalledWith(expect.stringContaining(".cache"), {
+        recursive: true,
+      });
     });
   });
 });
diff --git a/src/pricing.ts b/src/pricing.ts
index a579f28..b7d9e0c 100644
--- a/src/pricing.ts
+++ b/src/pricing.ts
@@ -1,5 +1,5 @@
-import { readFileSync, writeFileSync, mkdirSync, existsSync } from "node:fs";
-import { join, dirname } from "node:path";
+import { existsSync, mkdirSync, readFileSync, writeFileSync } from "node:fs";
+import { dirname, join } from "node:path";
 import { fileURLToPath } from "node:url";
 
 const __dirname = dirname(fileURLToPath(import.meta.url));
@@ -40,10 +40,7 @@ interface CacheData {
 
 let cache: CacheData | null = null;
 
-function normalize(
-  key: string,
-  raw: Record<string, unknown>,
-): ModelEntry | null {
+function normalize(key: string, raw: Record<string, unknown>): ModelEntry | null {
   const inputCost = Number(raw.input_cost_per_token) || 0;
   const outputCost = Number(raw.output_cost_per_token) || 0;
 
@@ -57,9 +54,7 @@ function normalize(
       : null;
 
   const cacheReadCost =
-    typeof raw.cache_read_input_token_cost === "number"
-      ? raw.cache_read_input_token_cost
-      : null;
+    typeof raw.cache_read_input_token_cost === "number" ? raw.cache_read_input_token_cost : null;
 
   return {
     key,
@@ -69,26 +64,19 @@ function normalize(
     output_cost_per_million: outputCost * 1_000_000,
     input_cost_per_token_above_200k: tieredInput,
     output_cost_per_token_above_200k: tieredOutput,
-    input_cost_per_million_above_200k:
-      tieredInput != null ? tieredInput * 1_000_000 : null,
-    output_cost_per_million_above_200k:
-      tieredOutput != null ? tieredOutput * 1_000_000 : null,
+    input_cost_per_million_above_200k: tieredInput != null ? tieredInput * 1_000_000 : null,
+    output_cost_per_million_above_200k: tieredOutput != null ? tieredOutput * 1_000_000 : null,
     cache_read_input_token_cost: cacheReadCost,
     cache_read_input_token_cost_per_million:
       cacheReadCost != null ? cacheReadCost * 1_000_000 : null,
-    max_input_tokens:
-      typeof raw.max_input_tokens === "number" ? raw.max_input_tokens : null,
-    max_output_tokens:
-      typeof raw.max_output_tokens === "number" ? raw.max_output_tokens : null,
-    max_tokens:
-      typeof raw.max_tokens === "number" ? raw.max_tokens : null,
-    litellm_provider:
-      typeof raw.litellm_provider === "string" ? raw.litellm_provider : "unknown",
+    max_input_tokens: typeof raw.max_input_tokens === "number" ? raw.max_input_tokens : null,
+    max_output_tokens: typeof raw.max_output_tokens === "number" ? raw.max_output_tokens : null,
+    max_tokens: typeof raw.max_tokens === "number" ? raw.max_tokens : null,
+    litellm_provider: typeof raw.litellm_provider === "string" ? raw.litellm_provider : "unknown",
     mode: typeof raw.mode === "string" ? raw.mode : "chat",
     supports_vision: raw.supports_vision === true,
     supports_function_calling: raw.supports_function_calling === true,
-    supports_parallel_function_calling:
-      raw.supports_parallel_function_calling === true,
+    supports_parallel_function_calling: raw.supports_parallel_function_calling === true,
   };
 }
 
@@ -180,12 +168,13 @@ export async function getModels(): Promise<Record<string, ModelEntry>> {
     if (cache) {
       return cache.models;
     }
-    throw new Error(
-      "No pricing data available — fetch failed and no cache exists",
-    );
+    throw new Error("No pricing data available — fetch failed and no cache exists");
   }
 
-  return cache!.models;
+  if (!cache) {
+    throw new Error("No pricing data available — refresh succeeded but cache is empty");
+  }
+  return cache.models;
 }
 
 export function getModelCount(): number {
diff --git a/src/search.test.ts b/src/search.test.ts
index 740f356..a37ce3b 100644
--- a/src/search.test.ts
+++ b/src/search.test.ts
@@ -1,6 +1,6 @@
-import { describe, it, expect } from "vitest";
-import { fuzzyMatch, fuzzyMatchMultiple, fuzzyMatchWithMetadata } from "./search.js";
+import { describe, expect, it } from "vitest";
 import type { ModelEntry } from "./pricing.js";
+import { fuzzyMatch, fuzzyMatchMultiple, fuzzyMatchWithMetadata } from "./search.js";
 
 /** Helper to build a minimal ModelEntry for testing */
 function makeModel(overrides: Partial<ModelEntry> & { key: string }): ModelEntry {
@@ -76,92 +76,92 @@ describe("fuzzyMatch", () => {
   it("strips azure/ prefix before matching", () => {
     const result = fuzzyMatch("azure/gpt-4o", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("strips bedrock/ prefix before matching", () => {
     const result = fuzzyMatch("bedrock/gpt-4o", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("strips vertex_ai/ prefix before matching", () => {
     const result = fuzzyMatch("vertex_ai/gemini-2.0-flash", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gemini-2.0-flash");
+    expect(result?.key).toBe("gemini-2.0-flash");
   });
 
   it("strips vertex_ai_beta/ prefix before matching", () => {
     const result = fuzzyMatch("vertex_ai_beta/claude-sonnet-4-5", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("claude-sonnet-4-5");
+    expect(result?.key).toBe("claude-sonnet-4-5");
   });
 
   it("strips openrouter/ prefix before matching", () => {
     const result = fuzzyMatch("openrouter/gpt-4o-mini", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o-mini");
+    expect(result?.key).toBe("gpt-4o-mini");
   });
 
   it("strips together_ai/ prefix before matching", () => {
     const result = fuzzyMatch("together_ai/claude-opus-4", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("claude-opus-4");
+    expect(result?.key).toBe("claude-opus-4");
   });
 
   it("strips fireworks_ai/ prefix before matching", () => {
     const result = fuzzyMatch("fireworks_ai/gpt-4o", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("strips prefix with case-insensitive matching", () => {
     const result = fuzzyMatch("AZURE/GPT-4O", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("returns exact match when key matches exactly", () => {
     const result = fuzzyMatch("gpt-4o", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("returns exact match for full model name", () => {
     const result = fuzzyMatch("claude-sonnet-4-5", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("claude-sonnet-4-5");
+    expect(result?.key).toBe("claude-sonnet-4-5");
   });
 
   it("handles case-insensitive exact match", () => {
     const result = fuzzyMatch("GPT-4O", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("handles mixed case exact match", () => {
     const result = fuzzyMatch("Claude-Sonnet-4-5", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("claude-sonnet-4-5");
+    expect(result?.key).toBe("claude-sonnet-4-5");
   });
 
   it("fuzzy matches partial/close model names", () => {
     const result = fuzzyMatch("claude sonnet", sampleModels);
     expect(result).not.toBeNull();
     // Should match one of the claude models
-    expect(result!.key).toContain("claude");
+    expect(result?.key).toContain("claude");
   });
 
   it("fuzzy matches with typos", () => {
     const result = fuzzyMatch("gpt-4o-mni", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o-mini");
+    expect(result?.key).toBe("gpt-4o-mini");
   });
 
   it("fuzzy matches gemini models", () => {
     const result = fuzzyMatch("gemini-2.0", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toContain("gemini");
+    expect(result?.key).toContain("gemini");
   });
 
   it("returns null for garbage input", () => {
@@ -242,26 +242,26 @@ describe("fine-tuned model patterns (ft: prefix)", () => {
   it("extracts base model from OpenAI fine-tuned pattern", () => {
     const result = fuzzyMatch("ft:gpt-4o:my-org:custom_suffix:id", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("extracts base model from fine-tuned gpt-4o-mini", () => {
     const result = fuzzyMatch("ft:gpt-4o-mini:my-org:suffix:abc123", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o-mini");
+    expect(result?.key).toBe("gpt-4o-mini");
   });
 
   it("returns metadata indicating fine-tuned status", () => {
     const result = fuzzyMatchWithMetadata("ft:gpt-4o:my-org:custom:id", sampleModels);
     expect(result.entry).not.toBeNull();
-    expect(result.entry!.key).toBe("gpt-4o");
+    expect(result.entry?.key).toBe("gpt-4o");
     expect(result.isFineTuned).toBe(true);
   });
 
   it("returns metadata with isFineTuned=false for non-fine-tuned models", () => {
     const result = fuzzyMatchWithMetadata("gpt-4o", sampleModels);
     expect(result.entry).not.toBeNull();
-    expect(result.entry!.key).toBe("gpt-4o");
+    expect(result.entry?.key).toBe("gpt-4o");
     expect(result.isFineTuned).toBe(false);
   });
 
@@ -269,13 +269,13 @@ describe("fine-tuned model patterns (ft: prefix)", () => {
     // ft: should be processed before provider prefix
     const result = fuzzyMatch("azure/ft:gpt-4o:my-org:custom:id", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("handles case-insensitive fine-tuned pattern", () => {
     const result = fuzzyMatch("FT:GPT-4O:MY-ORG:CUSTOM:ID", sampleModels);
     expect(result).not.toBeNull();
-    expect(result!.key).toBe("gpt-4o");
+    expect(result?.key).toBe("gpt-4o");
   });
 
   it("fuzzyMatchMultiple handles fine-tuned patterns", () => {
diff --git a/src/search.ts b/src/search.ts
index 93c99a8..44c28bc 100644
--- a/src/search.ts
+++ b/src/search.ts
@@ -68,14 +68,10 @@ function buildIndex(models: Record<string, ModelEntry>): Fuse<{ key: string }> {
   return fuse;
 }
 
-export function fuzzyMatch(
-  query: string,
-  models: Record<string, ModelEntry>,
-): ModelEntry | null {
+export function fuzzyMatch(query: string, models: Record<string, ModelEntry>): ModelEntry | null {
   // Strip provider prefix first, then handle fine-tuned pattern
   const withoutProvider = stripProviderPrefix(query);
-  const { base, isFineTuned } = extractFineTunedBase(withoutProvider);
-  const normalizedQuery = base;
+  const { base: normalizedQuery } = extractFineTunedBase(withoutProvider);
 
   // Try exact match first
   if (models[normalizedQuery]) {
@@ -106,7 +102,7 @@ export function fuzzyMatchWithMetadata(
   models: Record<string, ModelEntry>,
 ): SearchResult {
   const withoutProvider = stripProviderPrefix(query);
-  const { base, isFineTuned } = extractFineTunedBase(withoutProvider);
+  const { isFineTuned } = extractFineTunedBase(withoutProvider);
   const entry = fuzzyMatch(query, models);
   return { entry, isFineTuned };
 }
diff --git a/src/tools.test.ts b/src/tools.test.ts
index 20942b1..2633677 100644
--- a/src/tools.test.ts
+++ b/src/tools.test.ts
@@ -1,4 +1,4 @@
-import { describe, it, expect, vi, beforeEach, afterEach } from "vitest";
+import { afterEach, beforeEach, describe, expect, it, vi } from "vitest";
 import type { ModelEntry } from "./pricing.js";
 
 // Mock the pricing module
@@ -11,9 +11,9 @@ vi.mock("./pricing.js", async (importOriginal) => {
   };
 });
 
+import { getModels, refreshPrices } from "./pricing.js";
 // Import after mock setup
-import { executeTool, calculateTieredCost } from "./tools.js";
-import { getModels, refreshPrices, TIERED_PRICING_THRESHOLD } from "./pricing.js";
+import { calculateTieredCost, executeTool } from "./tools.js";
 
 /** Helper to build a minimal ModelEntry */
 function makeModel(overrides: Partial<ModelEntry> & { key: string }): ModelEntry {
@@ -403,9 +403,7 @@ describe("executeTool", () => {
     });
 
     it("handles getModels throwing an error", async () => {
-      vi.mocked(getModels).mockRejectedValueOnce(
-        new Error("No pricing data available"),
-      );
+      vi.mocked(getModels).mockRejectedValueOnce(new Error("No pricing data available"));
 
       const result = await executeTool("get_model_details", {
         model_name: "gpt-4o",
@@ -417,9 +415,7 @@ describe("executeTool", () => {
     });
 
     it("handles refreshPrices throwing an error", async () => {
-      vi.mocked(refreshPrices).mockRejectedValueOnce(
-        new Error("Network failure"),
-      );
+      vi.mocked(refreshPrices).mockRejectedValueOnce(new Error("Network failure"));
 
       const result = await executeTool("refresh_prices", {});
 
diff --git a/src/tools.ts b/src/tools.ts
index f73a654..9955232 100644
--- a/src/tools.ts
+++ b/src/tools.ts
@@ -1,11 +1,6 @@
 import { z } from "zod";
-import {
-  getModels,
-  refreshPrices,
-  TIERED_PRICING_THRESHOLD,
-  type ModelEntry,
-} from "./pricing.js";
-import { fuzzyMatch, fuzzyMatchWithMetadata } from "./search.js";
+import { getModels, type ModelEntry, refreshPrices, TIERED_PRICING_THRESHOLD } from "./pricing.js";
+import { fuzzyMatchWithMetadata } from "./search.js";
 
 export const tools = [
   {
@@ -61,8 +56,7 @@ export const tools = [
       properties: {
         provider: {
           type: "string",
-          description:
-            "Filter by provider (e.g. 'anthropic', 'openai', 'google', 'amazon')",
+          description: "Filter by provider (e.g. 'anthropic', 'openai', 'google', 'amazon')",
         },
         min_context: {
           type: "number",
@@ -144,14 +138,10 @@ export function calculateTieredCost(
   let inputTieredCost = 0;
   let tieredInput = false;
 
-  if (
-    model.input_cost_per_token_above_200k != null &&
-    inputTokens > threshold
-  ) {
+  if (model.input_cost_per_token_above_200k != null && inputTokens > threshold) {
     tieredInput = true;
     inputBaseCost = threshold * model.input_cost_per_token;
-    inputTieredCost =
-      (inputTokens - threshold) * model.input_cost_per_token_above_200k;
+    inputTieredCost = (inputTokens - threshold) * model.input_cost_per_token_above_200k;
   } else {
     inputBaseCost = inputTokens * model.input_cost_per_token;
   }
@@ -160,14 +150,10 @@ export function calculateTieredCost(
   let outputTieredCost = 0;
   let tieredOutput = false;
 
-  if (
-    model.output_cost_per_token_above_200k != null &&
-    outputTokens > threshold
-  ) {
+  if (model.output_cost_per_token_above_200k != null && outputTokens > threshold) {
     tieredOutput = true;
     outputBaseCost = threshold * model.output_cost_per_token;
-    outputTieredCost =
-      (outputTokens - threshold) * model.output_cost_per_token_above_200k;
+    outputTieredCost = (outputTokens - threshold) * model.output_cost_per_token_above_200k;
   } else {
     outputBaseCost = outputTokens * model.output_cost_per_token;
   }
@@ -192,8 +178,7 @@ function formatModelDetails(model: ModelEntry): string {
   const capabilities: string[] = [];
   if (model.supports_vision) capabilities.push("vision");
   if (model.supports_function_calling) capabilities.push("function_calling");
-  if (model.supports_parallel_function_calling)
-    capabilities.push("parallel_function_calling");
+  if (model.supports_parallel_function_calling) capabilities.push("parallel_function_calling");
 
   const hasTieredInput = model.input_cost_per_million_above_200k != null;
   const hasTieredOutput = model.output_cost_per_million_above_200k != null;
@@ -206,14 +191,10 @@ function formatModelDetails(model: ModelEntry): string {
       `Tiered Pricing (above ${formatTokenCount(TIERED_PRICING_THRESHOLD)} tokens, per 1M):`,
     );
     if (hasTieredInput) {
-      tieredLines.push(
-        `  Input:  ${formatCost(model.input_cost_per_million_above_200k!)}`,
-      );
+      tieredLines.push(`  Input:  ${formatCost(model.input_cost_per_million_above_200k ?? 0)}`);
     }
     if (hasTieredOutput) {
-      tieredLines.push(
-        `  Output: ${formatCost(model.output_cost_per_million_above_200k!)}`,
-      );
+      tieredLines.push(`  Output: ${formatCost(model.output_cost_per_million_above_200k ?? 0)}`);
     }
   }
 
@@ -240,9 +221,7 @@ function formatModelDetails(model: ModelEntry): string {
     `Context Window:`,
     `  Max Input:  ${formatTokenCount(model.max_input_tokens)}`,
     `  Max Output: ${formatTokenCount(model.max_output_tokens)}`,
-    ...(model.max_tokens !== null
-      ? [`  Max Tokens: ${formatTokenCount(model.max_tokens)}`]
-      : []),
+    ...(model.max_tokens !== null ? [`  Max Tokens: ${formatTokenCount(model.max_tokens)}`] : []),
     ``,
     `Capabilities: ${capabilities.length > 0 ? capabilities.join(", ") : "none listed"}`,
   ].join("\n");
@@ -299,9 +278,7 @@ export async function executeTool(
 
         // Resolve cached token count: cap at input_tokens, ignore if model doesn't support caching
         const resolvedCachedTokens =
-          cached_tokens != null &&
-          model.cache_read_input_token_cost != null &&
-          cached_tokens > 0
+          cached_tokens != null && model.cache_read_input_token_cost != null && cached_tokens > 0
             ? Math.min(cached_tokens, input_tokens)
             : 0;
         const uncachedInputTokens = input_tokens - resolvedCachedTokens;
@@ -309,7 +286,7 @@ export async function executeTool(
         const result = calculateTieredCost(model, uncachedInputTokens, output_tokens);
         const cachedCost =
           resolvedCachedTokens > 0
-            ? resolvedCachedTokens * model.cache_read_input_token_cost!
+            ? resolvedCachedTokens * (model.cache_read_input_token_cost ?? 0)
             : 0;
         const totalCost = result.totalCost + cachedCost;
 
@@ -324,7 +301,7 @@ export async function executeTool(
 
         if (resolvedCachedTokens > 0) {
           lines.push(
-            `  Cached input:   ${formatTokenCount(resolvedCachedTokens)} tokens × ${formatCost(model.cache_read_input_token_cost_per_million!)}/1M = ${formatCost(cachedCost)}`,
+            `  Cached input:   ${formatTokenCount(resolvedCachedTokens)} tokens × ${formatCost(model.cache_read_input_token_cost_per_million ?? 0)}/1M = ${formatCost(cachedCost)}`,
           );
         }
 
@@ -335,7 +312,7 @@ export async function executeTool(
             `  Input (base):  ${formatTokenCount(baseTokens)} tokens × ${formatCost(model.input_cost_per_million)}/1M = ${formatCost(result.inputBaseCost)}`,
           );
           lines.push(
-            `  Input (>200K): ${formatTokenCount(tieredTokens)} tokens × ${formatCost(model.input_cost_per_million_above_200k!)}/1M = ${formatCost(result.inputTieredCost)}`,
+            `  Input (>200K): ${formatTokenCount(tieredTokens)} tokens × ${formatCost(model.input_cost_per_million_above_200k ?? 0)}/1M = ${formatCost(result.inputTieredCost)}`,
           );
         } else {
           lines.push(
@@ -350,7 +327,7 @@ export async function executeTool(
             `  Output (base):  ${formatTokenCount(baseTokens)} tokens × ${formatCost(model.output_cost_per_million)}/1M = ${formatCost(result.outputBaseCost)}`,
           );
           lines.push(
-            `  Output (>200K): ${formatTokenCount(tieredTokens)} tokens × ${formatCost(model.output_cost_per_million_above_200k!)}/1M = ${formatCost(result.outputTieredCost)}`,
+            `  Output (>200K): ${formatTokenCount(tieredTokens)} tokens × ${formatCost(model.output_cost_per_million_above_200k ?? 0)}/1M = ${formatCost(result.outputTieredCost)}`,
           );
         } else {
           lines.push(
@@ -372,8 +349,7 @@ export async function executeTool(
       }
 
       case "compare_models": {
-        const { provider, min_context, mode } =
-          compareModelsSchema.parse(args);
+        const { provider, min_context, mode } = compareModelsSchema.parse(args);
         const models = await getModels();
 
         let filtered = Object.values(models);
@@ -389,17 +365,13 @@ export async function executeTool(
 
         if (min_context !== undefined) {
           filtered = filtered.filter(
-            (m) =>
-              m.max_input_tokens !== null &&
-              m.max_input_tokens >= min_context,
+            (m) => m.max_input_tokens !== null && m.max_input_tokens >= min_context,
           );
         }
 
         if (mode) {
           const lowerMode = mode.toLowerCase();
-          filtered = filtered.filter(
-            (m) => m.mode.toLowerCase() === lowerMode,
-          );
+          filtered = filtered.filter((m) => m.mode.toLowerCase() === lowerMode);
         }
 
         if (filtered.length === 0) {
@@ -414,9 +386,7 @@ export async function executeTool(
         }
 
         // Sort by input cost (most cost-effective first)
-        filtered.sort(
-          (a, b) => a.input_cost_per_token - b.input_cost_per_token,
-        );
+        filtered.sort((a, b) => a.input_cost_per_token - b.input_cost_per_token);
         const top = filtered.slice(0, 5);
 
         const header = `Top ${top.length} most cost-effective models${provider ? ` (provider: ${provider})` : ""}${min_context ? ` (min context: ${formatTokenCount(min_context)})` : ""}${mode ? ` (mode: ${mode})` : ""}:\n`;
@@ -435,9 +405,7 @@ export async function executeTool(
         const total = `\n(${filtered.length} models matched total)`;
 
         return {
-          content: [
-            { type: "text", text: header + rows.join("\n\n") + total },
-          ],
+          content: [{ type: "text", text: header + rows.join("\n\n") + total }],
         };
       }
 
diff --git a/vitest.config.ts b/vitest.config.ts
index 4265d6d..5a95829 100644
--- a/vitest.config.ts
+++ b/vitest.config.ts
@@ -8,8 +8,14 @@ export default defineConfig({
     coverage: {
       provider: "v8",
       include: ["src/**/*.ts"],
-      exclude: ["src/**/*.test.ts"],
+      exclude: ["src/**/*.test.ts", "src/index.ts"],
       reporter: ["text", "json"],
+      thresholds: {
+        statements: 85,
+        branches: 80,
+        functions: 80,
+        lines: 85,
+      },
     },
   },
 });