diff --git a/chapters/01-sdk-first-request.md b/chapters/01-sdk-first-request.md
index a7b130b..508d800 100644
--- a/chapters/01-sdk-first-request.md
+++ b/chapters/01-sdk-first-request.md
@@ -6,18 +6,18 @@ Everything in this chapter is a single round trip: you send a JSON body, you get
 
 ```text
    your code
-      |
-      v
-  +-------------------------------------------------+
-  |  REQUEST   POST /v1/messages                    |
-  |    model, max_tokens, messages, system?         |
-  +-------------------------------------------------+
-      |
-      v
-  +-------------------------------------------------+
-  |  RESPONSE  Message                              |
-  |    content[], stop_reason, usage                |
-  +-------------------------------------------------+
+      │
+      ▼
+  ┌─────────────────────────────────────────────────┐
+  │  REQUEST   POST /v1/messages                    │
+  │    model, max_tokens, messages, system?         │
+  └─────────────────────────────────────────────────┘
+      │
+      ▼
+  ┌─────────────────────────────────────────────────┐
+  │  RESPONSE  Message                              │
+  │    content[], stop_reason, usage                │
+  └─────────────────────────────────────────────────┘
 ```
 
 Prerequisites: Bun installed and your API credentials in a `.env` file at the project root (the next section explains which variables). No earlier chapters required.
diff --git a/chapters/02-streaming.md b/chapters/02-streaming.md
index fe62be4..428566f 100644
--- a/chapters/02-streaming.md
+++ b/chapters/02-streaming.md
@@ -28,20 +28,20 @@ The streaming and blocking paths converge on one shape: a fully typed `Anthropic
 
 ```text
   message_start          usage.input_tokens
-      |
-      v
+      │
+      ▼
   content_block_start    index, type (e.g. text)
-      |
-      v
+      │
+      ▼
   content_block_delta    text_delta / input_json_delta   (repeats)
-      |
-      v
+      │
+      ▼
   content_block_stop     index
-      |
-      v
+      │
+      ▼
   message_delta          stop_reason + usage.output_tokens
-      |
-      v
+      │
+      ▼
   message_stop
 ```
 
diff --git a/chapters/03-repl-telegram.md b/chapters/03-repl-telegram.md
index 415ce64..24ddcc5 100644
--- a/chapters/03-repl-telegram.md
+++ b/chapters/03-repl-telegram.md
@@ -3,11 +3,10 @@
 💬 So far each script asked Claude one fixed question and exited. Now you'll hand the keyboard to a real person. A coding agent is, at heart, a loop: read input, build messages, call Claude, emit output, repeat. The Claude call stays the same; only the I/O around it changes - so in this chapter you wrap that one call in two different frontends, a terminal REPL and a Telegram bot.
 
 ```text
-  read input -> build messages -> call Claude -> emit output -> repeat
-                                          |
-                                          v
-                                          stop_reason 'end_turn'
-                                          ends the turn
+  read input → build messages → call Claude → emit output → repeat
+                                │
+                                ▼
+                                stop_reason 'end_turn' ends the turn
 ```
 
 You already have `new Anthropic()`, the env vars, and the content-array narrowing from Chapter 1, plus `stream()`, `.on('text')`, and `finalMessage()` from Chapter 2. Here you only add the input/output plumbing. The one new env var is `TELEGRAM_BOT_TOKEN`, which lives in `.env` next to your Anthropic credentials - Bun auto-loads it, and as always the key comes from the environment; never hardcode it.
diff --git a/chapters/05-tools.md b/chapters/05-tools.md
index 2b3f9dc..5229125 100644
--- a/chapters/05-tools.md
+++ b/chapters/05-tools.md
@@ -27,8 +27,19 @@ Notice that you find the call by narrowing `response.content` to the block whose
 Catching the call is half the round-trip; now you close it. The flow is a fixed five-step dance, and it is worth holding the whole shape in your head before you read the code:
 
 ```text
-create(tools) -> stop_reason 'tool_use' -> run the tool locally
-       -> create again with a tool_result -> final text answer
+       create() with tools
+                │
+                ▼
+     stop_reason: 'tool_use'
+                │
+                ▼
+      run the tool locally
+                │
+                ▼
+  create() with the tool_result
+                │
+                ▼
+        final text answer
 ```
 
 When `stop_reason === 'tool_use'`, you append the assistant's content **verbatim** as an assistant turn, run the tool yourself, then send a new `user` turn whose `content` is a `tool_result` block - an `Anthropic.ToolResultBlockParam` that echoes the `tool_use_id` exactly. That echo is the rule that binds request to answer: every `tool_use` block in a turn needs a matching `tool_result`, or the next `create` call rejects the array.