diff --git a/src/lib/usage-normalizer.ts b/src/lib/usage-normalizer.ts
new file mode 100644
index 00000000..e041878c
--- /dev/null
+++ b/src/lib/usage-normalizer.ts
@@ -0,0 +1,147 @@
+/**
+ * Pure (no DB, no network, no global state) helpers that convert upstream
+ * usage payloads from OpenAI and Anthropic into a single shape the rest of
+ * the system stores.
+ *
+ * Field rules (see docs/design/02-database-schema.md):
+ *   - Anthropic `cache_creation_input_tokens` is folded into `inputTokens`.
+ *   - OpenAI `completion_tokens` already includes reasoning tokens — never
+ *     add them on top of `outputTokens`. `reasoningTokens` is informational.
+ */
+
+export interface NormalizedUsage {
+  inputTokens: number
+  cachedInputTokens: number
+  outputTokens: number
+  reasoningTokens: number
+  totalTokens: number
+}
+
+export class UsageMissingError extends Error {
+  constructor(message = "Upstream stream never delivered usage information") {
+    super(message)
+    this.name = "UsageMissingError"
+  }
+}
+
+interface OpenAIUsageShape {
+  prompt_tokens?: number
+  completion_tokens?: number
+  total_tokens?: number
+  prompt_tokens_details?: { cached_tokens?: number }
+  completion_tokens_details?: { reasoning_tokens?: number }
+}
+
+interface AnthropicUsageShape {
+  input_tokens?: number
+  output_tokens?: number
+  cache_read_input_tokens?: number
+  cache_creation_input_tokens?: number
+}
+
+interface AnthropicMessageShape {
+  type?: string
+  message?: { usage?: AnthropicUsageShape }
+  usage?: AnthropicUsageShape
+}
+
+const numOr0 = (v: unknown): number => (typeof v === "number" ? v : 0)
+
+export function normalizeOpenAIFinal(usage: unknown): NormalizedUsage {
+  const u = (usage ?? {}) as OpenAIUsageShape
+  const inputTokens = numOr0(u.prompt_tokens)
+  const cachedInputTokens = numOr0(u.prompt_tokens_details?.cached_tokens)
+  const outputTokens = numOr0(u.completion_tokens)
+  const reasoningTokens = numOr0(u.completion_tokens_details?.reasoning_tokens)
+  const totalTokens = numOr0(u.total_tokens) || inputTokens + outputTokens
+  return {
+    inputTokens,
+    cachedInputTokens,
+    outputTokens,
+    reasoningTokens,
+    totalTokens,
+  }
+}
+
+export function normalizeAnthropicMessage(message: unknown): NormalizedUsage {
+  const m = (message ?? {}) as { usage?: AnthropicUsageShape }
+  const u = m.usage ?? {}
+  const baseInput = numOr0(u.input_tokens)
+  const cacheCreate = numOr0(u.cache_creation_input_tokens)
+  const cachedInputTokens = numOr0(u.cache_read_input_tokens)
+  const inputTokens = baseInput + cacheCreate
+  const outputTokens = numOr0(u.output_tokens)
+  return {
+    inputTokens,
+    cachedInputTokens,
+    outputTokens,
+    reasoningTokens: 0,
+    totalTokens: inputTokens + outputTokens,
+  }
+}
+
+export function normalizeEmbeddings(usage: unknown): NormalizedUsage {
+  const u = (usage ?? {}) as { prompt_tokens?: number; total_tokens?: number }
+  const inputTokens = numOr0(u.prompt_tokens)
+  return {
+    inputTokens,
+    cachedInputTokens: 0,
+    outputTokens: 0,
+    reasoningTokens: 0,
+    totalTokens: numOr0(u.total_tokens) || inputTokens,
+  }
+}
+
+export interface StreamUsageAccumulator {
+  feed(chunk: unknown): void
+  finalize(): NormalizedUsage
+}
+
+export function createOpenAIAccumulator(): StreamUsageAccumulator {
+  let saved: OpenAIUsageShape | undefined
+
+  return {
+    feed(chunk) {
+      const c = chunk as { usage?: OpenAIUsageShape } | null | undefined
+      if (c && c.usage) {
+        saved = c.usage
+      }
+    },
+    finalize() {
+      if (!saved) throw new UsageMissingError()
+      return normalizeOpenAIFinal(saved)
+    },
+  }
+}
+
+export function createAnthropicAccumulator(): StreamUsageAccumulator {
+  let inputTokens = 0
+  let cachedInputTokens = 0
+  let outputTokens = 0
+
+  return {
+    feed(chunk) {
+      const ev = (chunk ?? {}) as AnthropicMessageShape
+      if (ev.type === "message_start" && ev.message?.usage) {
+        const u = ev.message.usage
+        inputTokens =
+          numOr0(u.input_tokens) + numOr0(u.cache_creation_input_tokens)
+        cachedInputTokens = numOr0(u.cache_read_input_tokens)
+        outputTokens = numOr0(u.output_tokens)
+        return
+      }
+      if (ev.type === "message_delta" && ev.usage) {
+        outputTokens = Math.max(outputTokens, numOr0(ev.usage.output_tokens))
+      }
+    },
+    finalize() {
+      return {
+        inputTokens,
+        cachedInputTokens,
+        outputTokens,
+        reasoningTokens: 0,
+        totalTokens: inputTokens + outputTokens,
+      }
+    },
+  }
+}
diff --git a/tests/usage-normalizer.test.ts b/tests/usage-normalizer.test.ts
new file mode 100644
index 00000000..671f4e98
--- /dev/null
+++ b/tests/usage-normalizer.test.ts
@@ -0,0 +1,142 @@
+import { test, expect, describe } from "bun:test"
+
+import {
+  createAnthropicAccumulator,
+  createOpenAIAccumulator,
+  normalizeAnthropicMessage,
+  normalizeEmbeddings,
+  normalizeOpenAIFinal,
+  UsageMissingError,
+} from "../src/lib/usage-normalizer"
+
+describe("normalizeOpenAIFinal", () => {
+  test("maps prompt/completion/total + cached + reasoning", () => {
+    const out = normalizeOpenAIFinal({
+      prompt_tokens: 100,
+      completion_tokens: 50,
+      total_tokens: 150,
+      prompt_tokens_details: { cached_tokens: 20 },
+      completion_tokens_details: { reasoning_tokens: 10 },
+    })
+    expect(out).toEqual({
+      inputTokens: 100,
+      cachedInputTokens: 20,
+      outputTokens: 50,
+      reasoningTokens: 10,
+      totalTokens: 150,
+    })
+  })
+
+  test("missing fields default to 0", () => {
+    const out = normalizeOpenAIFinal({
+      prompt_tokens: 10,
+      completion_tokens: 5,
+    })
+    expect(out.inputTokens).toBe(10)
+    expect(out.outputTokens).toBe(5)
+    expect(out.cachedInputTokens).toBe(0)
+    expect(out.reasoningTokens).toBe(0)
+    expect(out.totalTokens).toBe(15)
+  })
+})
+
+describe("normalizeAnthropicMessage", () => {
+  test("folds cache_creation_input_tokens into inputTokens", () => {
+    const out = normalizeAnthropicMessage({
+      usage: {
+        input_tokens: 100,
+        cache_creation_input_tokens: 25,
+        cache_read_input_tokens: 75,
+        output_tokens: 30,
+      },
+    })
+    expect(out.inputTokens).toBe(125)
+    expect(out.cachedInputTokens).toBe(75)
+    expect(out.outputTokens).toBe(30)
+    expect(out.totalTokens).toBe(155)
+  })
+})
+
+describe("normalizeEmbeddings", () => {
+  test("uses prompt_tokens as input", () => {
+    const out = normalizeEmbeddings({ prompt_tokens: 12, total_tokens: 12 })
+    expect(out.inputTokens).toBe(12)
+    expect(out.outputTokens).toBe(0)
+    expect(out.totalTokens).toBe(12)
+  })
+})
+
+describe("createOpenAIAccumulator", () => {
+  test("captures usage from final chunk", () => {
+    const acc = createOpenAIAccumulator()
+    acc.feed({ choices: [{ delta: { content: "hi" } }] })
+    acc.feed({
+      choices: [],
+      usage: {
+        prompt_tokens: 10,
+        completion_tokens: 5,
+        total_tokens: 15,
+      },
+    })
+    const out = acc.finalize()
+    expect(out.inputTokens).toBe(10)
+    expect(out.outputTokens).toBe(5)
+  })
+
+  test("throws when usage chunk never arrives", () => {
+    const acc = createOpenAIAccumulator()
+    acc.feed({ choices: [{ delta: { content: "hi" } }] })
+    expect(() => acc.finalize()).toThrow(UsageMissingError)
+  })
+})
+
+describe("createAnthropicAccumulator", () => {
+  test("aggregates message_start + message_delta", () => {
+    const acc = createAnthropicAccumulator()
+    acc.feed({
+      type: "message_start",
+      message: {
+        usage: {
+          input_tokens: 50,
+          cache_creation_input_tokens: 10,
+          cache_read_input_tokens: 20,
+          output_tokens: 1,
+        },
+      },
+    })
+    acc.feed({
+      type: "content_block_delta",
+      delta: { type: "text_delta", text: "a" },
+    })
+    acc.feed({ type: "message_delta", usage: { output_tokens: 7 } })
+    acc.feed({ type: "message_delta", usage: { output_tokens: 12 } })
+    const out = acc.finalize()
+    expect(out.inputTokens).toBe(60)
+    expect(out.cachedInputTokens).toBe(20)
+    expect(out.outputTokens).toBe(12)
+    expect(out.totalTokens).toBe(72)
+  })
+
+  test("returns sane zeros if only message_start arrived", () => {
+    const acc = createAnthropicAccumulator()
+    acc.feed({
+      type: "message_start",
+      message: { usage: { input_tokens: 5, output_tokens: 1 } },
+    })
+    const out = acc.finalize()
+    expect(out.inputTokens).toBe(5)
+    expect(out.outputTokens).toBe(1)
+  })
+
+  test("returns zeros when nothing arrives", () => {
+    const acc = createAnthropicAccumulator()
+    const out = acc.finalize()
+    expect(out).toEqual({
+      inputTokens: 0,
+      cachedInputTokens: 0,
+      outputTokens: 0,
+      reasoningTokens: 0,
+      totalTokens: 0,
+    })
+  })
+})