clew-code 0.2.13 → 0.2.14

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (52) hide show
  1. package/dist/main.js +2596 -2527
  2. package/docs/architecture.html +148 -148
  3. package/docs/architecture.th.html +79 -79
  4. package/docs/clew-code-architecture.html +6 -6
  5. package/docs/commands.html +225 -224
  6. package/docs/commands.th.html +132 -131
  7. package/docs/configuration.html +147 -147
  8. package/docs/configuration.th.html +108 -108
  9. package/docs/daemon.html +129 -129
  10. package/docs/daemon.th.html +73 -73
  11. package/docs/features/bridge-mode.html +99 -99
  12. package/docs/features/bridge-mode.th.html +90 -90
  13. package/docs/features/evals.html +182 -182
  14. package/docs/features/evals.th.html +90 -90
  15. package/docs/features/peer.html +178 -178
  16. package/docs/features/searxng-search.html +151 -151
  17. package/docs/features/searxng-search.th.html +95 -95
  18. package/docs/features/sentry-setup.html +157 -157
  19. package/docs/features/sentry-setup.th.html +97 -97
  20. package/docs/index.html +298 -299
  21. package/docs/index.th.html +292 -292
  22. package/docs/installation.html +105 -105
  23. package/docs/installation.th.html +105 -105
  24. package/docs/internals/growthbook-ab-testing.html +113 -113
  25. package/docs/internals/growthbook-ab-testing.th.html +81 -81
  26. package/docs/internals/hidden-features.html +175 -149
  27. package/docs/internals/hidden-features.th.html +135 -109
  28. package/docs/loop.html +181 -181
  29. package/docs/loop.th.html +227 -227
  30. package/docs/mcp.html +247 -247
  31. package/docs/mcp.th.html +207 -207
  32. package/docs/models.html +110 -111
  33. package/docs/models.th.html +61 -61
  34. package/docs/peer.html +236 -236
  35. package/docs/peer.th.html +280 -280
  36. package/docs/permission-model.html +102 -102
  37. package/docs/permission-model.th.html +67 -67
  38. package/docs/plugins.html +102 -102
  39. package/docs/plugins.th.html +79 -79
  40. package/docs/providers.html +126 -126
  41. package/docs/providers.th.html +80 -80
  42. package/docs/quick-start.html +93 -93
  43. package/docs/quick-start.th.html +1 -1
  44. package/docs/research-memory.html +82 -82
  45. package/docs/research-memory.th.html +72 -72
  46. package/docs/skills.html +117 -117
  47. package/docs/skills.th.html +90 -90
  48. package/docs/tools.html +170 -170
  49. package/docs/tools.th.html +84 -84
  50. package/docs/troubleshooting.html +106 -106
  51. package/docs/troubleshooting.th.html +85 -85
  52. package/package.json +162 -164
package/docs/models.html CHANGED
@@ -1,111 +1,110 @@
1
- <!DOCTYPE html>
2
- <html lang="en">
3
- <head>
4
- <meta charset="UTF-8">
5
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
- <title>Models — Clew</title>
7
- <meta name="description" content="Model selection guide for Clew — how to pick, switch, and manage models across 27 providers.">
8
- <link rel="preconnect" href="https://fonts.googleapis.com">
9
- <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
10
- <link href="https://fonts.googleapis.com/css2?family=DM+Sans:wght@400;500;600;700&family=JetBrains+Mono:wght@400;500;600;700&display=swap" rel="stylesheet">
11
- <link rel="stylesheet" href="css/styles.css">
12
- <link rel="icon" type="image/svg+xml" href="./assets/clew.svg">
13
- </head>
14
- <body>
15
- <header class="header"></header>
16
- <div class="app"><aside class="sidebar" id="sidebar"></aside><div class="sidebar-overlay" id="sidebarOverlay"></div>
17
- <div class="content-wrap"><main class="content">
18
- <div class="breadcrumbs"><a href="index.html">Home</a><span class="sep">/</span><span>Models</span></div>
19
- <h1>Models</h1>
20
- <p class="section-subtitle">Clew supports 100+ models across 27 providers. Pick the right model for each task — from frontier reasoning to fast local inference.</p>
21
-
22
- <h2>Model Selection Guide</h2>
23
- <table>
24
- <tr><th>Workload</th><th>Recommended Models</th><th>Context</th><th>Notes</th></tr>
25
- <tr><td>Complex architecture &amp; large edits</td><td>claude-opus-4-7, claude-sonnet-4-6, gpt-5.5, deepseek-v4-pro</td><td>1M tokens</td><td>Best tool calling, strong reasoning, largest context</td></tr>
26
- <tr><td>Code review &amp; debugging</td><td>claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro</td><td>1M-2M tokens</td><td>Good instruction following, reliable output</td></tr>
27
- <tr><td>Quick search &amp; summarization</td><td>claude-haiku-4-5, deepseek-v4-flash, mistral-small-4, groq/llama-3.3-70b</td><td>128K-200K</td><td>Fast, cost-effective, lower latency</td></tr>
28
- <tr><td>Offline / air-gapped</td><td>ollama/llama4:70b, ollama/llama3.3</td><td>131K tokens</td><td>Fully local, no network needed</td></tr>
29
- <tr><td>Cost-sensitive bulk</td><td>kilocode/kilo-auto/free, opencode-go/minimax-m2.7, groq/llama-3.1-8b</td><td>Varies</td><td>Free or minimal cost tiers available</td></tr>
30
- <tr><td>Research &amp; reasoning</td><td>claude-opus-4-7, deepseek-v4-pro, gemini-3.1-pro</td><td>1M-2M tokens</td><td>Deep reasoning, chain-of-thought</td></tr>
31
- <tr><td>Vision / multimodal</td><td>claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro, deepseek-v4-flash</td><td>1M tokens</td><td>Image understanding, document analysis</td></tr>
32
- </table>
33
-
34
- <h2>Notable Models by Provider</h2>
35
- <table>
36
- <tr><th>Provider</th><th>Models Available</th></tr>
37
- <tr><td>Anthropic</td><td>claude-opus-4-7 (1M ctx), claude-sonnet-4-6 (1M ctx, recommended), claude-haiku-4-5 (200K ctx, fast)</td></tr>
38
- <tr><td>OpenAI</td><td>gpt-5.5 (1M ctx), gpt-5.4 (1M ctx), gpt-5.4-mini (fast)</td></tr>
39
- <tr><td>Google</td><td>gemini-3.1-pro (2M ctx), gemini-3.1-flash-lite (1M ctx, fast)</td></tr>
40
- <tr><td>DeepSeek</td><td>deepseek-v4-pro (1M ctx, MoE), deepseek-v4-flash (1M ctx, fast MoE)</td></tr>
41
- <tr><td>xAI</td><td>grok-4.3 (128K ctx)</td></tr>
42
- <tr><td>Mistral</td><td>mistral-medium-3.5, mistral-small-4, mistral-large-3</td></tr>
43
- <tr><td>Groq</td><td>llama-3.3-70b, llama-3.1-8b-instant (fast)</td></tr>
44
- <tr><td>Ollama</td><td>llama4:70b (local), llama3.3 (local), qwen3.6-plus (local)</td></tr>
45
- <tr><td>GitHub Copilot</td><td>claude-opus-4.7, claude-sonnet-4.6, claude-haiku-4.5, gpt-5.5</td></tr>
46
- <tr><td>OpenCode</td><td>claude-opus-4-7, claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro, kimi-k2.6</td></tr>
47
- <tr><td>KiloCode</td><td>kilo-auto/free, kilo-auto/balanced, kilo-auto/frontier</td></tr>
48
- <tr><td>Cohere</td><td>command-a-plus (128K ctx)</td></tr>
49
- <tr><td>Perplexity</td><td>sonar-pro, sonar, sonar-reasoning-pro</td></tr>
50
- <tr><td>Together AI</td><td>deepseek-v4, Qwen3.6, Llama-4</td></tr>
51
- <tr><td>NVIDIA NIM</td><td>deepseek-v4-pro, glm-5.1, nemotron-3</td></tr>
52
- <tr><td>Cerebras</td><td>qwen-3-235b (extremely fast)</td></tr>
53
- <tr><td>Moonshot (Kimi)</td><td>kimi-k2.6 (128K ctx)</td></tr>
54
- <tr><td>Zhipu (GLM)</td><td>glm-4.7 (128K ctx)</td></tr>
55
- <tr><td>Hugging Face</td><td>llama-3.3-70b-instruct</td></tr>
56
- <tr><td>Poe</td><td>claude-3-7-sonnet</td></tr>
57
- </table>
58
-
59
- <h2>Model Switching</h2>
60
- <pre><code>/model # Interactive picker (fetches live models from provider API when supported)
61
- /model list # List all available models from the active provider API
62
- /model claude-sonnet-4-7 # Switch by full ID
63
- /model sonnet # Alias-based switching
64
- --model opus # CLI flag at startup</code></pre>
65
- <p>For providers with a /v1/models endpoint (OpenAI, DeepSeek, NVIDIA, etc.), the picker fetches live models from the API instead of the static providers.json list.
66
- The picker shows <strong>all models from every provider</strong> in named sections — not just the active provider's models. Recent models appear at top with the active provider's default. Select any model from any provider directly.</p>
67
- <p>Model switching is instant — the next conversation turn uses the new model. Previous context is preserved.</p>
68
-
69
- <h2>Model Aliases</h2>
70
- <p>Short aliases are resolved to full model IDs: <code>sonnet</code>, <code>opus</code>, <code>haiku</code>, <code>gpt5</code>, <code>flash</code>, etc.</p>
71
-
72
- <h2>CLI Flags for Models</h2>
73
- <table>
74
- <tr><th>Flag</th><th>Description</th></tr>
75
- <tr><td><code>--model &lt;name&gt;</code></td><td>Set model at startup (e.g., <code>--model sonnet</code> or <code>--model claude-opus-4-7</code>)</td></tr>
76
- <tr><td><code>--effort &lt;level&gt;</code></td><td>Set reasoning effort: <code>low</code>, <code>medium</code>, <code>high</code>, <code>max</code></td></tr>
77
- <tr><td><code>--max-turns &lt;N&gt;</code></td><td>Limit agentic turns in non-interactive mode</td></tr>
78
- <tr><td><code>--thinking &lt;mode&gt;</code></td><td>Thinking mode: <code>enabled</code>, <code>adaptive</code>, <code>disabled</code></td></tr>
79
- <tr><td><code>--fallback-model &lt;model&gt;</code></td><td>Fallback when primary model is overloaded (print mode only)</td></tr>
80
- <tr><td><code>--task-budget &lt;tokens&gt;</code></td><td>API-side task budget in tokens</td></tr>
81
- <tr><td><code>--max-budget-usd &lt;amount&gt;</code></td><td>Maximum spend on API calls (print mode only)</td></tr>
82
- </table>
83
-
84
- <h2>Model Capabilities</h2>
85
- <p>Each model declaration in <code>providers.json</code> includes:</p>
86
- <ul>
87
- <li>Model ID and display label</li>
88
- <li>Context window (<code>maxContext</code>) and max output tokens (<code>maxOutput</code>)</li>
89
- <li>Tool calling type (<code>native</code>, <code>none</code>)</li>
90
- <li>Vision support</li>
91
- <li>Streaming mode (<code>full</code>, <code>partial</code>)</li>
92
- <li>Reasoning support</li>
93
- <li>System prompt support</li>
94
- <li>Tags (for filtering): <code>fast</code>, <code>verified</code>, <code>recommended</code>, <code>latest</code>, <code>free</code>, <code>vision</code>, <code>tools</code>, <code>native</code>, <code>moe</code>, <code>local</code></li>
95
- </ul>
96
-
97
- <footer class="footer">
98
- <span>Clew v0.2.7 — Open Source</span>
99
- <div class="footer-links">
100
- <a href="https://github.com/ClewCode/ClewCode">GitHub</a>
101
- <a href="https://github.com/ClewCode/ClewCode/issues">Issues</a>
102
- </div>
103
- </footer>
104
- </main>
105
- <nav class="toc-sidebar"></nav>
106
- </div>
107
- </div>
108
- <script src="js/main.js"></script>
109
- </body>
110
- </html>
111
-
1
+ <!DOCTYPE html>
2
+ <html lang="en">
3
+ <head>
4
+ <meta charset="UTF-8">
5
+ <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
+ <title>Models — Clew</title>
7
+ <meta name="description" content="Model selection guide for Clew — how to pick, switch, and manage models across 28 providers.">
8
+ <link rel="preconnect" href="https://fonts.googleapis.com">
9
+ <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
10
+ <link href="https://fonts.googleapis.com/css2?family=DM+Sans:wght@400;500;600;700&family=JetBrains+Mono:wght@400;500;600;700&display=swap" rel="stylesheet">
11
+ <link rel="stylesheet" href="css/styles.css">
12
+ <link rel="icon" type="image/svg+xml" href="./assets/clew.svg">
13
+ </head>
14
+ <body>
15
+ <header class="header"></header>
16
+ <div class="app"><aside class="sidebar" id="sidebar"></aside><div class="sidebar-overlay" id="sidebarOverlay"></div>
17
+ <div class="content-wrap"><main class="content">
18
+ <div class="breadcrumbs"><a href="index.html">Home</a><span class="sep">/</span><span>Models</span></div>
19
+ <h1>Models</h1>
20
+ <p class="section-subtitle">Clew supports 100+ models across 28 providers. Pick the right model for each task — from frontier reasoning to fast local inference.</p>
21
+
22
+ <h2>Model Selection Guide</h2>
23
+ <table>
24
+ <tr><th>Workload</th><th>Recommended Models</th><th>Context</th><th>Notes</th></tr>
25
+ <tr><td>Complex architecture &amp; large edits</td><td>claude-opus-4-7, claude-sonnet-4-6, gpt-5.5, deepseek-v4-pro</td><td>1M tokens</td><td>Best tool calling, strong reasoning, largest context</td></tr>
26
+ <tr><td>Code review &amp; debugging</td><td>claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro-preview</td><td>1M-2M tokens</td><td>Good instruction following, reliable output</td></tr>
27
+ <tr><td>Quick search &amp; summarization</td><td>claude-haiku-4-5, deepseek-v4-flash, mistral-small-4, groq/llama-3.3-70b</td><td>128K-200K</td><td>Fast, cost-effective, lower latency</td></tr>
28
+ <tr><td>Offline / air-gapped</td><td>ollama/llama4:70b, ollama/llama3.3</td><td>131K tokens</td><td>Fully local, no network needed</td></tr>
29
+ <tr><td>Cost-sensitive bulk</td><td>kilocode/kilo-auto/free, opencode-go/minimax-m2.7, groq/llama-3.1-8b</td><td>Varies</td><td>Free or minimal cost tiers available</td></tr>
30
+ <tr><td>Research &amp; reasoning</td><td>claude-opus-4-7, deepseek-v4-pro, gemini-3.1-pro-preview</td><td>1M-2M tokens</td><td>Deep reasoning, chain-of-thought</td></tr>
31
+ <tr><td>Vision / multimodal</td><td>claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro-preview, deepseek-v4-flash</td><td>1M tokens</td><td>Image understanding, document analysis</td></tr>
32
+ </table>
33
+
34
+ <h2>Notable Models by Provider</h2>
35
+ <table>
36
+ <tr><th>Provider</th><th>Models Available</th></tr>
37
+ <tr><td>Anthropic</td><td>claude-opus-4-7 (1M ctx), claude-sonnet-4-6 (1M ctx, recommended), claude-haiku-4-5 (200K ctx, fast)</td></tr>
38
+ <tr><td>OpenAI</td><td>gpt-5.5 (1M ctx), gpt-5.4 (1M ctx), gpt-5.4-mini (fast)</td></tr>
39
+ <tr><td>Google</td><td>gemini-3.1-pro-preview (2M ctx), gemini-3.1-flash-lite (1M ctx, fast)</td></tr>
40
+ <tr><td>DeepSeek</td><td>deepseek-v4-pro (1M ctx, MoE), deepseek-v4-flash (1M ctx, fast MoE)</td></tr>
41
+ <tr><td>xAI</td><td>grok-4.3 (128K ctx)</td></tr>
42
+ <tr><td>Mistral</td><td>mistral-medium-3.5, mistral-small-4, mistral-large-3</td></tr>
43
+ <tr><td>Groq</td><td>llama-3.3-70b, llama-3.1-8b-instant (fast)</td></tr>
44
+ <tr><td>Ollama</td><td>llama4:70b (local), llama3.3 (local), qwen3.6-plus (local)</td></tr>
45
+ <tr><td>OpenCode</td><td>claude-opus-4-7, claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro, kimi-k2.6</td></tr>
46
+ <tr><td>KiloCode</td><td>kilo-auto/free, kilo-auto/balanced, kilo-auto/frontier</td></tr>
47
+ <tr><td>Cohere</td><td>command-a-plus (128K ctx)</td></tr>
48
+ <tr><td>Perplexity</td><td>sonar-pro, sonar, sonar-reasoning-pro</td></tr>
49
+ <tr><td>Together AI</td><td>deepseek-v4, Qwen3.6, Llama-4</td></tr>
50
+ <tr><td>NVIDIA NIM</td><td>deepseek-v4-pro, glm-5.1, nemotron-3</td></tr>
51
+ <tr><td>Cerebras</td><td>qwen-3-235b (extremely fast)</td></tr>
52
+ <tr><td>Moonshot (Kimi)</td><td>kimi-k2.6 (128K ctx)</td></tr>
53
+ <tr><td>Zhipu (GLM)</td><td>glm-4.7 (128K ctx)</td></tr>
54
+ <tr><td>Hugging Face</td><td>llama-3.3-70b-instruct</td></tr>
55
+ <tr><td>Poe</td><td>claude-3-7-sonnet</td></tr>
56
+ </table>
57
+
58
+ <h2>Model Switching</h2>
59
+ <pre><code>/model # Interactive picker (fetches live models from provider API when supported)
60
+ /model list # List all available models from the active provider API
61
+ /model claude-sonnet-4-7 # Switch by full ID
62
+ /model sonnet # Alias-based switching
63
+ --model opus # CLI flag at startup</code></pre>
64
+ <p>For providers with a /v1/models endpoint (OpenAI, DeepSeek, NVIDIA, etc.), the picker fetches live models from the API instead of the static providers.json list.
65
+ The picker shows <strong>all models from every provider</strong> in named sections — not just the active provider's models. Recent models appear at top with the active provider's default. Select any model from any provider directly.</p>
66
+ <p>Model switching is instant — the next conversation turn uses the new model. Previous context is preserved.</p>
67
+
68
+ <h2>Model Aliases</h2>
69
+ <p>Short aliases are resolved to full model IDs: <code>sonnet</code>, <code>opus</code>, <code>haiku</code>, <code>gpt5</code>, <code>flash</code>, etc.</p>
70
+
71
+ <h2>CLI Flags for Models</h2>
72
+ <table>
73
+ <tr><th>Flag</th><th>Description</th></tr>
74
+ <tr><td><code>--model &lt;name&gt;</code></td><td>Set model at startup (e.g., <code>--model sonnet</code> or <code>--model claude-opus-4-7</code>)</td></tr>
75
+ <tr><td><code>--effort &lt;level&gt;</code></td><td>Set reasoning effort: <code>low</code>, <code>medium</code>, <code>high</code>, <code>max</code></td></tr>
76
+ <tr><td><code>--max-turns &lt;N&gt;</code></td><td>Limit agentic turns in non-interactive mode</td></tr>
77
+ <tr><td><code>--thinking &lt;mode&gt;</code></td><td>Thinking mode: <code>enabled</code>, <code>adaptive</code>, <code>disabled</code></td></tr>
78
+ <tr><td><code>--fallback-model &lt;model&gt;</code></td><td>Fallback when primary model is overloaded (print mode only)</td></tr>
79
+ <tr><td><code>--task-budget &lt;tokens&gt;</code></td><td>API-side task budget in tokens</td></tr>
80
+ <tr><td><code>--max-budget-usd &lt;amount&gt;</code></td><td>Maximum spend on API calls (print mode only)</td></tr>
81
+ </table>
82
+
83
+ <h2>Model Capabilities</h2>
84
+ <p>Each model declaration in <code>providers.json</code> includes:</p>
85
+ <ul>
86
+ <li>Model ID and display label</li>
87
+ <li>Context window (<code>maxContext</code>) and max output tokens (<code>maxOutput</code>)</li>
88
+ <li>Tool calling type (<code>native</code>, <code>none</code>)</li>
89
+ <li>Vision support</li>
90
+ <li>Streaming mode (<code>full</code>, <code>partial</code>)</li>
91
+ <li>Reasoning support</li>
92
+ <li>System prompt support</li>
93
+ <li>Tags (for filtering): <code>fast</code>, <code>verified</code>, <code>recommended</code>, <code>latest</code>, <code>free</code>, <code>vision</code>, <code>tools</code>, <code>native</code>, <code>moe</code>, <code>local</code></li>
94
+ </ul>
95
+
96
+ <footer class="footer">
97
+ <span>Clew 0.2.13 — Open Source</span>
98
+ <div class="footer-links">
99
+ <a href="https://github.com/ClewCode/ClewCode">GitHub</a>
100
+ <a href="https://github.com/ClewCode/ClewCode/issues">Issues</a>
101
+ </div>
102
+ </footer>
103
+ </main>
104
+ <nav class="toc-sidebar"></nav>
105
+ </div>
106
+ </div>
107
+ <script src="js/main.js"></script>
108
+ </body>
109
+ </html>
110
+
package/docs/models.th.html CHANGED
@@ -1,61 +1,61 @@
1
- <!DOCTYPE html>
2
- <html lang="th">
3
- <head>
4
- <meta charset="UTF-8">
5
- <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
- <title>โมเดล — Clew</title>
7
- <meta name="description" content="คู่มือการเลือกโมเดลสำหรับ Clew — วิธีเลือก, สลับ, และจัดการโมเดลจาก 27 ผู้ให้บริการ">
8
- <link rel="preconnect" href="https://fonts.googleapis.com">
9
- <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
10
- <link href="https://fonts.googleapis.com/css2?family=DM+Sans:wght@400;500;600;700&family=JetBrains+Mono:wght@400;500;600;700&family=Noto+Sans+Thai:wght@400;500;600;700&display=swap" rel="stylesheet">
11
- <link rel="stylesheet" href="css/styles.css">
12
- <link rel="icon" type="image/svg+xml" href="./assets/clew.svg">
13
- </head>
14
- <body>
15
- <header class="header"></header>
16
- <div class="app"><aside class="sidebar" id="sidebar"></aside><div class="sidebar-overlay" id="sidebarOverlay"></div>
17
- <div class="content-wrap"><main class="content">
18
- <div class="breadcrumbs"><a href="index.th.html">หน้าแรก</a><span class="sep">/</span><span>โมเดล</span></div>
19
- <h1>โมเดล</h1>
20
- <p class="section-subtitle">Clew รองรับ 100+ โมเดลจาก 27 ผู้ให้บริการ เลือกโมเดลที่เหมาะกับแต่ละงาน — จากการใช้เหตุผลระดับสูงไปจนถึงการอนุมานในเครื่องที่รวดเร็ว</p>
21
-
22
- <h2>คู่มือการเลือกโมเดล</h2>
23
- <table>
24
- <tr><th>ลักษณะงาน</th><th>โมเดลแนะนำ</th><th>บริบท</th><th>หมายเหตุ</th></tr>
25
- <tr><td>สถาปัตยกรรมซับซ้อนและแก้ไขขนาดใหญ่</td><td>claude-opus-4-7, claude-sonnet-4-6, gpt-5.5, deepseek-v4-pro</td><td>1M โทเค็น</td><td>การเรียกใช้เครื่องมือดีที่สุด, การใช้เหตุผลแข็งแกร่ง, บริบทใหญ่ที่สุด</td></tr>
26
- <tr><td>รีวิวโค้ดและดีบัก</td><td>claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro</td><td>1M-2M โทเค็น</td><td>ทำตามคำแนะนำดี, ผลลัพธ์เชื่อถือได้</td></tr>
27
- <tr><td>ค้นหาและสรุปผลอย่างรวดเร็ว</td><td>claude-haiku-4-5, deepseek-v4-flash, mistral-small-4, groq/llama-3.3-70b</td><td>128K-200K</td><td>เร็ว, คุ้มค่า, หน่วงเวลาต่ำ</td></tr>
28
- <tr><td>ออฟไลน์ / ไม่มีเครือข่าย</td><td>ollama/llama4:70b, ollama/llama3.3</td><td>131K โทเค็น</td><td>ในเครื่องทั้งหมด, ไม่ต้องใช้เครือข่าย</td></tr>
29
- </table>
30
-
31
- <h2>การสลับโมเดล</h2>
32
- <pre><code>/model # ตัวเลือกแบบโต้ตอบ
33
- /model list # แสดงโมเดลทั้งหมดจากผู้ให้บริการที่ใช้งาน
34
- /model claude-sonnet-4-7 # สลับด้วย ID เต็ม
35
- /model sonnet # สลับด้วย alias
36
- --model opus # CLI flag ตอนเริ่ม</code></pre>
37
-
38
- <h2>CLI Flags สำหรับโมเดล</h2>
39
- <table>
40
- <tr><th>Flag</th><th>คำอธิบาย</th></tr>
41
- <tr><td><code>--model &lt;name&gt;</code></td><td>ตั้งค่าโมเดลตอนเริ่ม</td></tr>
42
- <tr><td><code>--effort &lt;level&gt;</code></td><td>ตั้งค่าระดับการใช้เหตุผล: low, medium, high, max</td></tr>
43
- <tr><td><code>--thinking &lt;mode&gt;</code></td><td>โหมดการคิด: enabled, adaptive, disabled</td></tr>
44
- <tr><td><code>--fallback-model &lt;model&gt;</code></td><td>โมเดลสำรองเมื่อโมเดลหลักโอเวอร์โหลด</td></tr>
45
- </table>
46
-
47
- <footer class="footer">
48
- <span>Clew v0.2.7 — โอเพนซอร์ส</span>
49
- <div class="footer-links">
50
- <a href="https://github.com/ClewCode/ClewCode">GitHub</a>
51
- <a href="https://github.com/ClewCode/ClewCode/issues">ปัญหา</a>
52
- </div>
53
- </footer>
54
- </main>
55
- <nav class="toc-sidebar"></nav>
56
- </div>
57
- </div>
58
- <script src="js/main.js"></script>
59
- </body>
60
- </html>
61
-
1
+ <!DOCTYPE html>
2
+ <html lang="th">
3
+ <head>
4
+ <meta charset="UTF-8">
5
+ <meta name="viewport" content="width=device-width, initial-scale=1.0">
6
+ <title>โมเดล — Clew</title>
7
+ <meta name="description" content="คู่มือการเลือกโมเดลสำหรับ Clew — วิธีเลือก, สลับ, และจัดการโมเดลจาก 27 ผู้ให้บริการ">
8
+ <link rel="preconnect" href="https://fonts.googleapis.com">
9
+ <link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
10
+ <link href="https://fonts.googleapis.com/css2?family=DM+Sans:wght@400;500;600;700&family=JetBrains+Mono:wght@400;500;600;700&family=Noto+Sans+Thai:wght@400;500;600;700&display=swap" rel="stylesheet">
11
+ <link rel="stylesheet" href="css/styles.css">
12
+ <link rel="icon" type="image/svg+xml" href="./assets/clew.svg">
13
+ </head>
14
+ <body>
15
+ <header class="header"></header>
16
+ <div class="app"><aside class="sidebar" id="sidebar"></aside><div class="sidebar-overlay" id="sidebarOverlay"></div>
17
+ <div class="content-wrap"><main class="content">
18
+ <div class="breadcrumbs"><a href="index.th.html">หน้าแรก</a><span class="sep">/</span><span>โมเดล</span></div>
19
+ <h1>โมเดล</h1>
20
+ <p class="section-subtitle">Clew รองรับ 100+ โมเดลจาก 27 ผู้ให้บริการ เลือกโมเดลที่เหมาะกับแต่ละงาน — จากการใช้เหตุผลระดับสูงไปจนถึงการอนุมานในเครื่องที่รวดเร็ว</p>
21
+
22
+ <h2>คู่มือการเลือกโมเดล</h2>
23
+ <table>
24
+ <tr><th>ลักษณะงาน</th><th>โมเดลแนะนำ</th><th>บริบท</th><th>หมายเหตุ</th></tr>
25
+ <tr><td>สถาปัตยกรรมซับซ้อนและแก้ไขขนาดใหญ่</td><td>claude-opus-4-7, claude-sonnet-4-6, gpt-5.5, deepseek-v4-pro</td><td>1M โทเค็น</td><td>การเรียกใช้เครื่องมือดีที่สุด, การใช้เหตุผลแข็งแกร่ง, บริบทใหญ่ที่สุด</td></tr>
26
+ <tr><td>รีวิวโค้ดและดีบัก</td><td>claude-sonnet-4-6, gpt-5.5, gemini-3.1-pro</td><td>1M-2M โทเค็น</td><td>ทำตามคำแนะนำดี, ผลลัพธ์เชื่อถือได้</td></tr>
27
+ <tr><td>ค้นหาและสรุปผลอย่างรวดเร็ว</td><td>claude-haiku-4-5, deepseek-v4-flash, mistral-small-4, groq/llama-3.3-70b</td><td>128K-200K</td><td>เร็ว, คุ้มค่า, หน่วงเวลาต่ำ</td></tr>
28
+ <tr><td>ออฟไลน์ / ไม่มีเครือข่าย</td><td>ollama/llama4:70b, ollama/llama3.3</td><td>131K โทเค็น</td><td>ในเครื่องทั้งหมด, ไม่ต้องใช้เครือข่าย</td></tr>
29
+ </table>
30
+
31
+ <h2>การสลับโมเดล</h2>
32
+ <pre><code>/model # ตัวเลือกแบบโต้ตอบ
33
+ /model list # แสดงโมเดลทั้งหมดจากผู้ให้บริการที่ใช้งาน
34
+ /model claude-sonnet-4-7 # สลับด้วย ID เต็ม
35
+ /model sonnet # สลับด้วย alias
36
+ --model opus # CLI flag ตอนเริ่ม</code></pre>
37
+
38
+ <h2>CLI Flags สำหรับโมเดล</h2>
39
+ <table>
40
+ <tr><th>Flag</th><th>คำอธิบาย</th></tr>
41
+ <tr><td><code>--model &lt;name&gt;</code></td><td>ตั้งค่าโมเดลตอนเริ่ม</td></tr>
42
+ <tr><td><code>--effort &lt;level&gt;</code></td><td>ตั้งค่าระดับการใช้เหตุผล: low, medium, high, max</td></tr>
43
+ <tr><td><code>--thinking &lt;mode&gt;</code></td><td>โหมดการคิด: enabled, adaptive, disabled</td></tr>
44
+ <tr><td><code>--fallback-model &lt;model&gt;</code></td><td>โมเดลสำรองเมื่อโมเดลหลักโอเวอร์โหลด</td></tr>
45
+ </table>
46
+
47
+ <footer class="footer">
48
+ <span>Clew 0.2.13 — โอเพนซอร์ส</span>
49
+ <div class="footer-links">
50
+ <a href="https://github.com/ClewCode/ClewCode">GitHub</a>
51
+ <a href="https://github.com/ClewCode/ClewCode/issues">ปัญหา</a>
52
+ </div>
53
+ </footer>
54
+ </main>
55
+ <nav class="toc-sidebar"></nav>
56
+ </div>
57
+ </div>
58
+ <script src="js/main.js"></script>
59
+ </body>
60
+ </html>
61
+