kward 0.85.0 → 0.86.0

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 (151) hide show
  1. checksums.yaml +4 -4
  2. data/CHANGELOG.md +130 -0
  3. data/CONTRIBUTING.md +1 -1
  4. data/Gemfile.lock +6 -2
  5. data/README.md +3 -3
  6. data/Rakefile +3 -0
  7. data/doc/agent-tools.md +11 -0
  8. data/doc/api.md +4 -1
  9. data/doc/composer.md +4 -3
  10. data/doc/configuration.md +60 -11
  11. data/doc/extensibility.md +2 -0
  12. data/doc/getting-started.md +3 -3
  13. data/doc/git.md +1 -1
  14. data/doc/memory.md +1 -1
  15. data/doc/plugins.md +40 -5
  16. data/doc/providers.md +10 -4
  17. data/doc/rpc.md +125 -7
  18. data/doc/session-management.md +36 -1
  19. data/doc/shell.md +643 -27
  20. data/doc/tabs.md +2 -1
  21. data/doc/troubleshooting.md +1 -1
  22. data/doc/usage.md +21 -11
  23. data/examples/plugins/openrouter_proxy/README.md +103 -0
  24. data/examples/plugins/openrouter_proxy/fake_embeddings.rb +85 -0
  25. data/examples/plugins/openrouter_proxy/plugin.rb +7 -0
  26. data/examples/plugins/openrouter_proxy/protocol.rb +202 -0
  27. data/examples/plugins/openrouter_proxy/server.rb +186 -0
  28. data/examples/plugins/space_invaders.rb +0 -1
  29. data/exe/kward +2 -2
  30. data/exe/kwsh +7 -0
  31. data/kward.gemspec +3 -1
  32. data/lib/kward/agent.rb +43 -9
  33. data/lib/kward/cli/commands.rb +41 -11
  34. data/lib/kward/cli/interactive_turn.rb +10 -0
  35. data/lib/kward/cli/plugins.rb +5 -1
  36. data/lib/kward/cli/prompt_interface.rb +4 -1
  37. data/lib/kward/cli/rendering.rb +21 -28
  38. data/lib/kward/cli/runtime_helpers.rb +52 -34
  39. data/lib/kward/cli/settings/model.rb +1 -12
  40. data/lib/kward/cli/tabs.rb +91 -18
  41. data/lib/kward/cli/tool_summaries.rb +2 -0
  42. data/lib/kward/cli/transcript_formatter.rb +3 -10
  43. data/lib/kward/cli/worktrees.rb +1 -0
  44. data/lib/kward/cli.rb +56 -7
  45. data/lib/kward/config/core.rb +16 -5
  46. data/lib/kward/config/extensions.rb +3 -0
  47. data/lib/kward/config/paths.rb +11 -0
  48. data/lib/kward/config/settings.rb +7 -0
  49. data/lib/kward/conversation.rb +11 -1
  50. data/lib/kward/entrypoint.rb +14 -0
  51. data/lib/kward/events.rb +2 -0
  52. data/lib/kward/generated_images.rb +237 -0
  53. data/lib/kward/model/catalog.rb +6 -3
  54. data/lib/kward/model/client.rb +105 -36
  55. data/lib/kward/model/codex_images.rb +83 -0
  56. data/lib/kward/model/copilot_models.rb +8 -4
  57. data/lib/kward/model/model_info.rb +27 -9
  58. data/lib/kward/model/openrouter_model_cache.rb +4 -2
  59. data/lib/kward/model/payloads.rb +6 -5
  60. data/lib/kward/model/sources.rb +3 -1
  61. data/lib/kward/model/stream_parser.rb +4 -7
  62. data/lib/kward/pan/index.html.erb +8 -1
  63. data/lib/kward/pan/server.rb +13 -1
  64. data/lib/kward/plugins/codex_responses.rb +73 -0
  65. data/lib/kward/plugins/host.rb +16 -0
  66. data/lib/kward/plugins/http_server.rb +83 -0
  67. data/lib/kward/plugins/registry.rb +39 -0
  68. data/lib/kward/prompt_interface/composer_controller.rb +0 -1
  69. data/lib/kward/prompt_interface/composer_renderer.rb +159 -72
  70. data/lib/kward/prompt_interface/editor/controller.rb +1 -1
  71. data/lib/kward/prompt_interface/editor/renderer.rb +10 -7
  72. data/lib/kward/prompt_interface/file_overlay.rb +3 -3
  73. data/lib/kward/prompt_interface/generated_image_preview.rb +162 -0
  74. data/lib/kward/prompt_interface/git_prompt.rb +18 -11
  75. data/lib/kward/prompt_interface/interactive/renderer.rb +1 -1
  76. data/lib/kward/prompt_interface/interactive/state.rb +18 -20
  77. data/lib/kward/prompt_interface/key_handler.rb +108 -25
  78. data/lib/kward/prompt_interface/layout.rb +11 -1
  79. data/lib/kward/prompt_interface/overlay_renderer.rb +25 -8
  80. data/lib/kward/prompt_interface/project_browser.rb +27 -54
  81. data/lib/kward/prompt_interface/prompt_renderer.rb +27 -35
  82. data/lib/kward/prompt_interface/question_prompt.rb +19 -28
  83. data/lib/kward/prompt_interface/runtime_state.rb +31 -9
  84. data/lib/kward/prompt_interface/screen.rb +24 -21
  85. data/lib/kward/prompt_interface/selection_prompt.rb +8 -8
  86. data/lib/kward/prompt_interface/slash_overlay.rb +1 -1
  87. data/lib/kward/prompt_interface/transcript_renderer.rb +7 -6
  88. data/lib/kward/prompt_interface.rb +109 -39
  89. data/lib/kward/prompts.rb +7 -1
  90. data/lib/kward/pty/adaptive_output_sink.rb +38 -4
  91. data/lib/kward/pty/interactive_runner.rb +20 -28
  92. data/lib/kward/pty/transcript_normalizer.rb +65 -21
  93. data/lib/kward/rpc/config_manager.rb +3 -0
  94. data/lib/kward/rpc/server.rb +27 -5
  95. data/lib/kward/rpc/session_manager.rb +62 -17
  96. data/lib/kward/rpc/transcript_normalizer.rb +11 -7
  97. data/lib/kward/sessions/store.rb +30 -3
  98. data/lib/kward/shell/kwsh.rb +27 -7
  99. data/lib/kward/shell/kwshrc.rb +11 -2
  100. data/lib/kward/shell/native_arithmetic.rb +255 -0
  101. data/lib/kward/shell/native_arrays.rb +102 -0
  102. data/lib/kward/shell/native_braces.rb +96 -0
  103. data/lib/kward/shell/native_builtins.rb +476 -0
  104. data/lib/kward/shell/native_completion.rb +169 -0
  105. data/lib/kward/shell/native_condition.rb +75 -0
  106. data/lib/kward/shell/native_directory.rb +116 -0
  107. data/lib/kward/shell/native_exec_state.rb +73 -0
  108. data/lib/kward/shell/native_expansion.rb +252 -0
  109. data/lib/kward/shell/native_fields.rb +62 -0
  110. data/lib/kward/shell/native_glob.rb +96 -0
  111. data/lib/kward/shell/native_history.rb +242 -0
  112. data/lib/kward/shell/native_interactive.rb +164 -0
  113. data/lib/kward/shell/native_invocation.rb +42 -0
  114. data/lib/kward/shell/native_jobs.rb +250 -0
  115. data/lib/kward/shell/native_line_editor.rb +382 -0
  116. data/lib/kward/shell/native_line_terminal.rb +286 -0
  117. data/lib/kward/shell/native_locale.rb +109 -0
  118. data/lib/kward/shell/native_locals.rb +57 -0
  119. data/lib/kward/shell/native_pattern.rb +200 -0
  120. data/lib/kward/shell/native_process.rb +279 -0
  121. data/lib/kward/shell/native_program.rb +40 -0
  122. data/lib/kward/shell/native_programmable_completion.rb +115 -0
  123. data/lib/kward/shell/native_prompt_format.rb +89 -0
  124. data/lib/kward/shell/native_runtime.rb +964 -0
  125. data/lib/kward/shell/native_substitution.rb +110 -0
  126. data/lib/kward/shell/native_syntax.rb +946 -0
  127. data/lib/kward/shell/native_traps.rb +196 -0
  128. data/lib/kward/shell/native_vi.rb +407 -0
  129. data/lib/kward/shell/native_vi_terminal.rb +19 -0
  130. data/lib/kward/shell/native_worker.rb +82 -0
  131. data/lib/kward/shell/persistent_session.rb +184 -190
  132. data/lib/kward/shell/prompt.rb +14 -0
  133. data/lib/kward/shell/prompt_agent.rb +17 -0
  134. data/lib/kward/shell/script_runner.rb +244 -0
  135. data/lib/kward/shell/standalone_agent.rb +31 -0
  136. data/lib/kward/shell/standalone_link.rb +78 -0
  137. data/lib/kward/shell/standalone_terminal.rb +287 -0
  138. data/lib/kward/terminal/keys.rb +22 -1
  139. data/lib/kward/terminal/locale_text.rb +29 -0
  140. data/lib/kward/terminal/sequences.rb +45 -4
  141. data/lib/kward/tools/datetime.rb +85 -0
  142. data/lib/kward/tools/generate_image.rb +36 -0
  143. data/lib/kward/tools/registry.rb +35 -9
  144. data/lib/kward/tools/tool_call.rb +3 -1
  145. data/lib/kward/transcripts/markdown_transcript.rb +8 -0
  146. data/lib/kward/transcripts/turn_summary_formatter.rb +93 -0
  147. data/lib/kward/transport/gateway.rb +2 -1
  148. data/lib/kward/turn_summary.rb +68 -0
  149. data/lib/kward/version.rb +1 -1
  150. data/lib/main.rb +6 -3
  151. metadata +84 -1
@@ -15,6 +15,7 @@ require_relative "copilot_models"
15
15
  require_relative "model_info"
16
16
  require_relative "sources"
17
17
  require_relative "payloads"
18
+ require_relative "codex_images"
18
19
  require_relative "../telemetry/logger"
19
20
  require_relative "stream_parser"
20
21
 
@@ -110,7 +111,7 @@ module Kward
110
111
  @local_models = nil
111
112
  end
112
113
 
113
- def chat(messages, tools: [], on_reasoning_delta: nil, on_reasoning_boundary: nil, on_assistant_delta: nil, on_retry: nil, cancellation: nil, steering: nil, max_tokens: nil, provider: nil, model: nil, reasoning: nil, provider_required: false)
114
+ def chat(messages, tools: [], on_reasoning_delta: nil, on_reasoning_boundary: nil, on_assistant_delta: nil, on_retry: nil, cancellation: nil, steering: nil, max_tokens: nil, provider: nil, model: nil, reasoning: nil, provider_required: false, text_format: nil)
114
115
  cancellation&.raise_if_cancelled!
115
116
  requested_provider = provider
116
117
  validate_required_provider!(requested_provider) if provider_required
@@ -126,7 +127,8 @@ module Kward
126
127
  current_model = resolved_copilot_chat_model(current_model) if resolved_provider == "Copilot" && model.nil?
127
128
 
128
129
  validate_image_support!(resolved_provider, current_model, messages)
129
- request_body = JSON.dump(request_body_payload(resolved_provider, messages, tools, max_tokens: max_tokens, model: current_model, reasoning: reasoning))
130
+ tools = tools.reject { |tool| ToolCall.name(tool) == "generate_image" } if resolved_provider != "Codex"
131
+ request_body = JSON.dump(request_body_payload(resolved_provider, messages, tools, max_tokens: max_tokens, model: current_model, reasoning: reasoning, text_format: text_format))
130
132
  with_retries(resolved_provider, current_model, request_bytes: request_body.bytesize, on_retry: on_retry, cancellation: cancellation) do
131
133
  request_started_at = @telemetry_logger.monotonic_now
132
134
  message = nil
@@ -166,6 +168,16 @@ module Kward
166
168
  raise e
167
169
  end
168
170
 
171
+ # Generates via the Codex subscription only, never using a public API key.
172
+ def generate_codex_image(prompt, turn_id: nil, cancellation: nil)
173
+ _url, token, provider, account_id = credentials(provider: "Codex")
174
+ unless provider == "Codex" && !token.to_s.empty?
175
+ raise CodexImages::Error, "Sign in to Codex to generate images."
176
+ end
177
+
178
+ CodexImages.generate(prompt, token: token, account_id: account_id, turn_id: turn_id, cancellation: cancellation)
179
+ end
180
+
169
181
  # Returns the active provider label after applying env/config/credential fallback rules.
170
182
  def current_provider
171
183
  _url, _token, provider = credentials
@@ -226,11 +238,12 @@ module Kward
226
238
  metadata = openrouter_cached_model_entries.find { |entry| (entry["id"] || entry[:id]).to_s == id }
227
239
  model_entry("OpenRouter", id, current: provider == "OpenRouter" && openrouter_model == id, metadata: metadata)
228
240
  end
241
+ models << model_entry("OpenRouter", openrouter_model, current: provider == "OpenRouter") unless openrouter_model.to_s.empty? || openrouter_choices.include?(openrouter_model)
229
242
  end
230
243
 
231
244
  if provider_logged_in?("Copilot")
232
245
  copilot_model = model_for("Copilot")
233
- copilot_choices = provider == "Copilot" ? copilot_model_choices : static_copilot_model_choices
246
+ copilot_choices = provider == "Copilot" || @copilot_models ? copilot_model_choices : static_copilot_model_choices
234
247
  models += copilot_choices.map do |id|
235
248
  model_entry("Copilot", id, current: provider == "Copilot" && copilot_model == id)
236
249
  end
@@ -239,10 +252,13 @@ module Kward
239
252
 
240
253
  if provider_logged_in?("Anthropic")
241
254
  anthropic_model = model_for("Anthropic")
242
- models += ModelInfo::ANTHROPIC_MODEL_CHOICES.map do |id|
243
- model_entry("Anthropic", id, current: provider == "Anthropic" && anthropic_model == id)
255
+ anthropic_entries = ModelSources.new(provider_id: "anthropic", api_key: @api_key_store.fetch("anthropic")).models
256
+ anthropic_choices = (anthropic_entries.map { |entry| entry.fetch("id") } + ModelInfo::ANTHROPIC_MODEL_CHOICES).uniq
257
+ models += anthropic_choices.map do |id|
258
+ metadata = anthropic_entries.find { |entry| entry.fetch("id") == id }
259
+ model_entry("Anthropic", id, current: provider == "Anthropic" && anthropic_model == id, metadata: metadata)
244
260
  end
245
- models << model_entry("Anthropic", anthropic_model, current: provider == "Anthropic") unless ModelInfo::ANTHROPIC_MODEL_CHOICES.include?(anthropic_model)
261
+ models << model_entry("Anthropic", anthropic_model, current: provider == "Anthropic") unless anthropic_choices.include?(anthropic_model)
246
262
  end
247
263
 
248
264
  ProviderCatalog.api_key_providers.each do |catalog_provider|
@@ -270,15 +286,43 @@ module Kward
270
286
  models.sort_by { |model| [model[:provider], model[:id]] }
271
287
  end
272
288
 
273
- # Refreshes cached choices for an API-key provider.
289
+ # Retains the array return value used by terminal pickers and internal callers.
274
290
  def refresh_available_models(provider: nil)
291
+ refresh_model_catalog(provider: provider)
292
+ available_models
293
+ end
294
+
295
+ # Refreshes only one provider and reports a safe, frontend-neutral outcome.
296
+ # Failed discovery never replaces the last usable catalog or returns raw errors.
297
+ def refresh_model_catalog(provider: nil)
275
298
  provider ||= current_provider
299
+ provider = "OpenAI" if provider.to_s == "openai_api"
276
300
  catalog_provider = ProviderCatalog.find_by_name(provider) || ProviderCatalog.find(provider.to_s.downcase)
277
- return available_models unless catalog_provider&.api_key?
278
- return available_models unless ProviderCatalog.runtime(catalog_provider.id).automatic_model_discovery?
301
+ provider = ModelInfo.provider_label(catalog_provider&.id == "openai" ? "OpenAI" : catalog_provider&.id || provider)
302
+ status = "success"
279
303
 
280
- ModelSources.new(provider_id: catalog_provider.id, api_key: @api_key_store.fetch(catalog_provider.id)).refresh
281
- available_models
304
+ case provider
305
+ when "OpenRouter"
306
+ @openrouter_models = openrouter_model_cache.refresh.fetch("models")
307
+ when "Copilot"
308
+ fetch_copilot_models(refresh: true)
309
+ when "Local"
310
+ local_model_choices(refresh: true)
311
+ else
312
+ if catalog_provider&.api_key? && ProviderCatalog.runtime(catalog_provider.id).automatic_model_discovery?
313
+ ModelSources.new(provider_id: catalog_provider.id, api_key: @api_key_store.fetch(catalog_provider.id)).refresh(strict: true)
314
+ else
315
+ status = "unsupported"
316
+ end
317
+ end
318
+
319
+ { provider: provider, status: status, fallbackUsed: status != "success" }
320
+ rescue StandardError
321
+ # Avoid an implicit retry during the following list read when no prior
322
+ # in-memory catalog exists. Configured/curated choices remain available.
323
+ @copilot_models ||= [] if provider == "Copilot"
324
+ @local_models ||= [] if provider == "Local"
325
+ { provider: provider, status: "failed", fallbackUsed: true }
282
326
  end
283
327
 
284
328
  # Projects messages/tools into the provider-specific context shape without sending it.
@@ -691,7 +735,7 @@ module Kward
691
735
  end
692
736
  end
693
737
 
694
- def request_body_payload(provider, messages, tools, max_tokens: nil, model: nil, reasoning: nil)
738
+ def request_body_payload(provider, messages, tools, max_tokens: nil, model: nil, reasoning: nil, text_format: nil)
695
739
  reasoning = false unless ModelInfo.reasoning_supported?(provider, model)
696
740
 
697
741
  if provider == "Azure OpenAI"
@@ -699,9 +743,9 @@ module Kward
699
743
  elsif provider == "Google Gemini"
700
744
  gemini_payload(messages, tools, max_tokens: max_tokens, model: model)
701
745
  elsif provider == "OpenAI"
702
- openai_responses_payload(messages, tools, max_tokens: max_tokens, model: model, reasoning: reasoning)
746
+ openai_responses_payload(messages, tools, max_tokens: max_tokens, model: model, reasoning: reasoning, text_format: text_format)
703
747
  elsif provider == "Codex"
704
- codex_payload(messages, tools, max_tokens: max_tokens, model: model, reasoning: reasoning)
748
+ codex_payload(messages, tools, max_tokens: max_tokens, model: model, reasoning: reasoning, text_format: text_format)
705
749
  elsif provider == "Anthropic"
706
750
  anthropic_payload(messages, tools, max_tokens: max_tokens, model: model, reasoning: reasoning)
707
751
  elsif provider == "Copilot" && copilot_responses_model?(model)
@@ -733,23 +777,27 @@ module Kward
733
777
  ModelInfo.context_window(provider, id, openrouter_models: openrouter_cached_model_entries)
734
778
  end
735
779
 
736
- def local_model_choices
737
- return @local_models if @local_models
780
+ def local_model_choices(refresh: false)
781
+ return @local_models if @local_models && !refresh
738
782
 
739
783
  url = local_models_url
740
784
  request = Http.apply_user_agent(Net::HTTP::Get.new(url))
741
785
  request["Authorization"] = "Bearer #{local_api_key}" unless local_api_key.to_s.empty?
742
786
  request["Accept"] = "application/json"
743
787
  response = Net::HTTP.start(url.hostname, url.port, use_ssl: url.scheme == "https", read_timeout: stream_idle_timeout_seconds) { |http| http.request(request) }
744
- return @local_models = [] unless response.is_a?(Net::HTTPSuccess)
788
+ raise "Local model discovery failed" unless response.is_a?(Net::HTTPSuccess)
745
789
 
746
790
  body = JSON.parse(response.body)
747
- entries = body["data"] || body["models"] || []
748
- @local_models = Array(entries).filter_map do |entry|
791
+ entries = body.is_a?(Hash) && (body["data"] || body["models"])
792
+ raise "Invalid Local model catalog" unless entries.is_a?(Array)
793
+
794
+ @local_models = entries.filter_map do |entry|
749
795
  entry.is_a?(Hash) ? entry["id"] || entry[:id] || entry["model"] || entry[:model] : entry
750
796
  end.map(&:to_s).map(&:strip).reject(&:empty?).uniq
751
797
  rescue StandardError
752
- @local_models = []
798
+ raise if refresh
799
+
800
+ @local_models ||= []
753
801
  end
754
802
 
755
803
  def openrouter_model_choices
@@ -787,11 +835,11 @@ module Kward
787
835
  CopilotModels.resolved_chat_model(configured_model, fetch_copilot_models)
788
836
  end
789
837
 
790
- def fetch_copilot_models
791
- return @copilot_models if @copilot_models
838
+ def fetch_copilot_models(refresh: false)
839
+ return @copilot_models if @copilot_models && !refresh
792
840
 
793
841
  token = github_access_token.to_s
794
- return [] if token.empty?
842
+ raise "No Copilot token found" if token.empty?
795
843
 
796
844
  url = URI("#{@github_oauth.base_url}/models")
797
845
  request = Http.apply_user_agent(Net::HTTP::Get.new(url))
@@ -800,10 +848,12 @@ module Kward
800
848
  copilot_headers([]).each { |key, value| request[key] = value }
801
849
 
802
850
  response = Net::HTTP.start(url.hostname, url.port, use_ssl: true) { |http| http.request(request) }
803
- return [] unless response.is_a?(Net::HTTPSuccess)
851
+ raise "Copilot model discovery failed" unless response.is_a?(Net::HTTPSuccess)
804
852
 
805
- @copilot_models = CopilotModels.parse(response.body)
853
+ @copilot_models = CopilotModels.parse(response.body, strict: refresh)
806
854
  rescue StandardError
855
+ raise if refresh
856
+
807
857
  []
808
858
  end
809
859
 
@@ -972,24 +1022,38 @@ module Kward
972
1022
 
973
1023
  input_tokens = integer_value(usage, "input_tokens", "prompt_tokens")
974
1024
  output_tokens = integer_value(usage, "output_tokens", "completion_tokens")
975
- cache_read_tokens = positive_integer(
1025
+ cache_read_tokens = token_count(
976
1026
  nested_value(usage, "input_tokens_details", "cached_tokens") ||
977
1027
  nested_value(usage, "prompt_tokens_details", "cached_tokens") ||
978
1028
  usage["cache_read_tokens"] || usage[:cache_read_tokens] || usage["cacheReadTokens"] || usage[:cacheReadTokens] || usage["cache_read_input_tokens"] || usage[:cache_read_input_tokens]
979
1029
  )
980
- cache_write_tokens = integer_value(usage, "cache_write_tokens", "cacheWriteTokens", "cache_creation_input_tokens")
1030
+ cache_write_tokens = token_count(
1031
+ nested_value(usage, "input_tokens_details", "cache_write_tokens") ||
1032
+ integer_value(usage, "cache_write_tokens", "cacheWriteTokens", "cache_creation_input_tokens")
1033
+ )
1034
+ reasoning_tokens = token_count(
1035
+ nested_value(usage, "output_tokens_details", "reasoning_tokens") ||
1036
+ nested_value(usage, "completion_tokens_details", "reasoning_tokens") ||
1037
+ integer_value(usage, "reasoning_tokens")
1038
+ )
1039
+ # Anthropic reports uncached input separately; OpenAI/Gemini include cache
1040
+ # tokens in input. Normalize to inclusive input before computing totals.
1041
+ if input_tokens && %w[cache_read_input_tokens cache_creation_input_tokens].any? { |key| usage.key?(key) || usage.key?(key.to_sym) }
1042
+ input_tokens += cache_read_tokens.to_i + cache_write_tokens.to_i
1043
+ end
981
1044
  total_tokens = integer_value(usage, "total_tokens", "totalTokens")
982
- total_tokens ||= [input_tokens, output_tokens, cache_read_tokens, cache_write_tokens].compact.sum
983
- return nil unless total_tokens&.positive? || input_tokens&.positive? || output_tokens&.positive?
1045
+ total_tokens ||= input_tokens.to_i + output_tokens.to_i if input_tokens || output_tokens
1046
+ return nil if total_tokens.nil?
984
1047
 
985
1048
  {
986
- "input_tokens" => input_tokens || 0,
987
- "output_tokens" => output_tokens || 0,
988
- "cache_read_tokens" => cache_read_tokens || 0,
989
- "cache_write_tokens" => cache_write_tokens || 0,
990
- "total_tokens" => total_tokens || 0,
1049
+ "input_tokens" => input_tokens,
1050
+ "output_tokens" => output_tokens,
1051
+ "cache_read_tokens" => cache_read_tokens,
1052
+ "cache_write_tokens" => cache_write_tokens,
1053
+ "reasoning_tokens" => reasoning_tokens,
1054
+ "total_tokens" => total_tokens,
991
1055
  "estimated" => false
992
- }
1056
+ }.compact
993
1057
  end
994
1058
 
995
1059
  def integer_value(source, *keys)
@@ -998,7 +1062,7 @@ module Kward
998
1062
  key = keys.find { |candidate| source.key?(candidate) || source.key?(candidate.to_sym) }
999
1063
  return nil unless key
1000
1064
 
1001
- positive_integer(source[key] || source[key.to_sym])
1065
+ token_count(source[key] || source[key.to_sym])
1002
1066
  end
1003
1067
 
1004
1068
  def nested_value(source, outer_key, inner_key)
@@ -1008,6 +1072,11 @@ module Kward
1008
1072
  outer[inner_key] || outer[inner_key.to_sym]
1009
1073
  end
1010
1074
 
1075
+ def token_count(value)
1076
+ integer = Integer(value, exception: false)
1077
+ integer if integer && integer >= 0
1078
+ end
1079
+
1011
1080
  def positive_integer(value)
1012
1081
  integer = value.to_i
1013
1082
  integer.positive? ? integer : nil
@@ -0,0 +1,83 @@
1
+ require "base64"
2
+ require "json"
3
+ require "net/http"
4
+ require "uri"
5
+ require_relative "../cancellation"
6
+ require_relative "../http"
7
+
8
+ module Kward
9
+ # Codex subscription transport, separate from the chat Responses stream.
10
+ class CodexImages
11
+ URL = URI("https://chatgpt.com/backend-api/codex/images/generations")
12
+ MAX_IMAGE_BYTES = 20 * 1024 * 1024
13
+ MAX_RESPONSE_BYTES = (MAX_IMAGE_BYTES + 2) / 3 * 4 + 65_536
14
+ PNG_HEADER = "\x89PNG\r\n\x1a\n\x00\x00\x00\x0dIHDR".b.freeze
15
+ Error = Class.new(StandardError)
16
+
17
+ def self.generate(prompt, token:, account_id:, turn_id:, cancellation: nil)
18
+ request = Http.apply_user_agent(Net::HTTP::Post.new(URL))
19
+ request["Authorization"] = "Bearer #{token}"
20
+ request["ChatGPT-Account-Id"] = account_id unless account_id.to_s.empty?
21
+ request["x-codex-image-turn-id"] = turn_id unless turn_id.to_s.empty?
22
+ request["originator"] = "kward"
23
+ request["Content-Type"] = "application/json"
24
+ request["Accept"] = "application/json"
25
+ request.body = JSON.generate(prompt: prompt, model: "gpt-image-2", background: "opaque", quality: "auto", size: "auto")
26
+
27
+ body = +""
28
+ cancellation&.raise_if_cancelled!
29
+ Net::HTTP.start(URL.hostname, URL.port, use_ssl: true, open_timeout: 10, read_timeout: 180) do |http|
30
+ http.max_retries = 0
31
+ cancellation&.on_cancel { http.finish if http.started? }
32
+ cancellation&.raise_if_cancelled!
33
+ http.request(request) do |response|
34
+ raise Error, error_message(response.code.to_i) unless response.is_a?(Net::HTTPSuccess)
35
+
36
+ response.read_body do |chunk|
37
+ cancellation&.raise_if_cancelled!
38
+ raise Error, "Codex image exceeds the preview size limit." if body.bytesize + chunk.bytesize > MAX_RESPONSE_BYTES
39
+
40
+ body << chunk
41
+ end
42
+ end
43
+ end
44
+ cancellation&.raise_if_cancelled!
45
+ decode(body)
46
+ rescue IOError, SystemCallError, SocketError, Timeout::Error, OpenSSL::SSL::SSLError, Net::HTTPBadResponse
47
+ cancellation&.raise_if_cancelled!
48
+ raise Error, "The connection to Codex image generation was interrupted; it was not retried."
49
+ end
50
+
51
+ def self.decode(body)
52
+ result = JSON.parse(body)
53
+ data = result.is_a?(Hash) ? result["data"] : nil
54
+ first = data.first if data.is_a?(Array)
55
+ encoded = first["b64_json"] if first.is_a?(Hash)
56
+ raise Error, "Codex did not return an image." unless encoded.is_a?(String)
57
+
58
+ bytes = Base64.strict_decode64(encoded)
59
+ raise Error, "Codex returned an invalid or oversized PNG." unless valid_png?(bytes)
60
+
61
+ bytes
62
+ rescue JSON::ParserError, ArgumentError
63
+ raise Error, "Codex returned an invalid image response."
64
+ end
65
+
66
+ def self.valid_png?(bytes)
67
+ return false unless bytes.is_a?(String) && bytes.bytesize.between?(45, MAX_IMAGE_BYTES) && bytes.b.start_with?(PNG_HEADER)
68
+
69
+ width, height = bytes.byteslice(16, 8).unpack("NN")
70
+ width.between?(1, 8192) && height.between?(1, 8192) && bytes.end_with?("\x00\x00\x00\x00IEND\xaeB`\x82".b)
71
+ end
72
+
73
+ def self.error_message(status)
74
+ case status
75
+ when 401 then "Codex sign-in expired; sign in again to generate an image."
76
+ when 403, 404 then "Image generation is not available for this Codex account."
77
+ when 429 then "Codex image generation is rate-limited or the account has reached its usage limit."
78
+ else "Codex could not generate an image (HTTP #{status})."
79
+ end
80
+ end
81
+ private_class_method :decode, :error_message
82
+ end
83
+ end
@@ -6,9 +6,11 @@ module Kward
6
6
  module CopilotModels
7
7
  module_function
8
8
 
9
- def parse(body)
10
- catalog_entries(body).filter_map { |entry| model_id(entry) }.uniq
9
+ def parse(body, strict: false)
10
+ catalog_entries(body, strict: strict).filter_map { |entry| model_id(entry) }.uniq
11
11
  rescue JSON::ParserError
12
+ raise if strict
13
+
12
14
  []
13
15
  end
14
16
 
@@ -27,9 +29,11 @@ module Kward
27
29
  choices.find { |model| supported?(model) } || raise("No Copilot models supported by Kward are available for this account. Kward currently supports Copilot GPT-5/GPT-6 Responses and Gemini/GPT-4.1 chat models.")
28
30
  end
29
31
 
30
- def catalog_entries(body)
32
+ def catalog_entries(body, strict: false)
31
33
  data = JSON.parse(body.to_s)
32
- entries = data.is_a?(Hash) ? data["data"] || data["models"] || data["items"] || [] : data
34
+ entries = data.is_a?(Hash) ? data["data"] || data["models"] || data["items"] : data
35
+ raise "Invalid Copilot model catalog" if strict && !entries.is_a?(Array)
36
+
33
37
  Array(entries)
34
38
  end
35
39
 
@@ -9,8 +9,13 @@ module Kward
9
9
  DEFAULT_COPILOT_MODEL = "gpt-5-mini"
10
10
  DEFAULT_ANTHROPIC_MODEL = "claude-sonnet-5"
11
11
  DEFAULT_REASONING_EFFORT = "medium"
12
- OPENAI_MODEL_CHOICES = %w[gpt-6-astra gpt-6.1-sol gpt-6-luna gpt-5.6-sol gpt-5.6-terra gpt-5.6-luna gpt-5.5 gpt-5.4 gpt-5.4-mini gpt-5.3-codex-spark].freeze
12
+ OPENAI_MODEL_CHOICES = %w[gpt-6.2-astra gpt-6.2-sol gpt-6.1-terra gpt-6-astra gpt-6.1-sol gpt-6-luna gpt-5.6-sol gpt-5.6-terra gpt-5.6-luna gpt-5.5 gpt-5.4 gpt-5.4-mini gpt-5.3-codex-spark].freeze
13
13
  ANTHROPIC_MODEL_CHOICES = %w[
14
+ claude-fable-5-2
15
+ claude-mythos-5-2
16
+ claude-opus-5-5
17
+ claude-sonnet-5-5
18
+ claude-haiku-5-5
14
19
  claude-fable-5
15
20
  claude-opus-5
16
21
  claude-opus-4-8
@@ -87,7 +92,7 @@ module Kward
87
92
  ].freeze
88
93
 
89
94
  CODEX_CONTEXT_WINDOWS = [
90
- [/\Agpt-6(?:\.1)?-(?:astra|sol|luna)(?:\z|-)/, 1_050_000],
95
+ [/\Agpt-6(?:\.\d+)?-(?:astra|sol|luna|terra)(?:\z|-)/, 1_050_000],
91
96
  [/\Agpt-5\.6/, 1_050_000],
92
97
  [/\Agpt-5\.5/, 400_000],
93
98
  [/\Agpt-5\.4-mini/, 400_000],
@@ -106,7 +111,7 @@ module Kward
106
111
  [/\Agpt-3\.5-turbo/, 16_385]
107
112
  ].freeze
108
113
  OPENAI_CONTEXT_WINDOWS = [
109
- [/\Agpt-6(?:\.1)?-(?:astra|sol|luna)(?:\z|-)/, 1_050_000],
114
+ [/\Agpt-6(?:\.\d+)?-(?:astra|sol|luna|terra)(?:\z|-)/, 1_050_000],
110
115
  [/\Agpt-5\.6/, 1_050_000],
111
116
  [/\Agpt-5\.5/, 1_050_000],
112
117
  [/\Agpt-5\.4-mini/, 400_000],
@@ -125,7 +130,7 @@ module Kward
125
130
  [/\Agpt-3\.5-turbo/, 16_385]
126
131
  ].freeze
127
132
  ANTHROPIC_CONTEXT_WINDOWS = [
128
- [/\Aclaude-(?:fable|mythos|opus|sonnet)-5(?:\z|-)/, 1_000_000],
133
+ [/\Aclaude-(?:fable|mythos|opus|sonnet|haiku)-5(?:\z|-)/, 1_000_000],
129
134
  [/\Aclaude-opus-4-(?:6|7|8)(?:\z|-)/, 1_000_000],
130
135
  [/\Aclaude-sonnet-4-6(?:\z|-)/, 1_000_000],
131
136
  [/\Aclaude-(?:haiku|opus|sonnet)-4-5(?:\z|-)/, 200_000],
@@ -435,7 +440,7 @@ module Kward
435
440
  when "Codex", "OpenRouter"
436
441
  true
437
442
  when "OpenAI"
438
- id.to_s.match?(/\Agpt-6(?:\.1)?-(?:astra|sol|luna)(?:\z|-)/)
443
+ id.to_s.match?(/\Agpt-6(?:\.\d+)?-(?:astra|sol|luna|terra)(?:\z|-)/)
439
444
  when "Anthropic"
440
445
  !reasoning_effort_choices(provider, id).empty?
441
446
  when "Copilot"
@@ -463,7 +468,8 @@ module Kward
463
468
  def openai_reasoning_effort_choices(id)
464
469
  text = id.to_s.delete_prefix("openai/")
465
470
  return GPT_6_LUNA_REASONING_EFFORT_CHOICES if text.match?(/\Agpt-6-luna(?:\z|-)/)
466
- return GPT_6_REASONING_EFFORT_CHOICES if text.match?(/\Agpt-6\.1-sol(?:\z|-)|\Agpt-6-astra(?:\z|-)/)
471
+ return OPENAI_MAX_REASONING_EFFORT_CHOICES if text.match?(/\Agpt-6\.1-terra(?:\z|-)/)
472
+ return GPT_6_REASONING_EFFORT_CHOICES if text.match?(/\Agpt-6\.2-(?:astra|sol)(?:\z|-)|\Agpt-6\.1-sol(?:\z|-)|\Agpt-6-astra(?:\z|-)/)
467
473
  return REASONING_EFFORT_CHOICES if text.match?(/\Agpt-5\.[23]-codex/)
468
474
  return OPENAI_MAX_REASONING_EFFORT_CHOICES if text.match?(/\Agpt-5\.6(?:\z|-)/)
469
475
 
@@ -472,13 +478,25 @@ module Kward
472
478
 
473
479
  def anthropic_reasoning_effort_choices(id)
474
480
  text = normalize_anthropic_model(id)
475
- return ANTHROPIC_HIGH_REASONING_EFFORT_CHOICES if text.match?(/\Aclaude-(?:fable|mythos|opus|sonnet)-5(?:\z|-)|\Aclaude-opus-4-(?:7|8)(?:\z|-)/)
481
+ return ANTHROPIC_HIGH_REASONING_EFFORT_CHOICES if text.match?(/\Aclaude-(?:fable|mythos|opus|sonnet|haiku)-5(?:\z|-)|\Aclaude-opus-4-(?:7|8)(?:\z|-)/)
476
482
  return ANTHROPIC_STANDARD_REASONING_EFFORT_CHOICES if text.match?(/\Aclaude-(?:opus-4-6|sonnet-4-6)(?:\z|-)/)
477
483
  return ANTHROPIC_OPUS_4_5_REASONING_EFFORT_CHOICES if text.match?(/\Aclaude-opus-4-5(?:\z|-)/)
478
484
 
479
485
  []
480
486
  end
481
487
 
488
+ # Unknown capabilities and the selection in the caller's context stay visible.
489
+ def default_selectable_model?(model)
490
+ return true if model[:current]
491
+
492
+ capabilities = Array(model[:supportedParameters]).map { |value| value.to_s.downcase }
493
+ return true if capabilities.empty?
494
+
495
+ capabilities.any? do |capability|
496
+ ["tools", "toolchoice", "functioncalling", "generatecontent", "streamgeneratecontent"].include?(capability.delete("_-"))
497
+ end
498
+ end
499
+
482
500
  def normalize(model, current_provider: nil, current_model: nil, current_reasoning_effort: nil)
483
501
  model = stringify_keys(model || {})
484
502
  provider = model["provider"]
@@ -494,8 +512,8 @@ module Kward
494
512
  contextWindow: model["contextWindow"] || model["context_window"] || context_window(provider, id),
495
513
  supportedParameters: model["supportedParameters"] || model["supported_parameters"],
496
514
  source: model["source"],
497
- current: boolean_value(model["current"], default: provider == current_provider && id == current_model)
498
- }.compact
515
+ current: current_provider && current_model ? provider == current_provider && id == current_model : boolean_value(model["current"])
516
+ }.compact.then { |entry| entry.merge(defaultSelectable: default_selectable_model?(entry)) }
499
517
  end
500
518
 
501
519
  def normalize_available(models, current_provider:, current_model:, current_reasoning_effort: nil)
@@ -80,8 +80,10 @@ module Kward
80
80
 
81
81
  def model_entries(body)
82
82
  data = JSON.parse(body.to_s)
83
- entries = data.is_a?(Hash) ? data["data"] || [] : data
84
- Array(entries).select { |entry| entry.is_a?(Hash) }
83
+ entries = data.is_a?(Hash) ? data["data"] : data
84
+ raise "Invalid OpenRouter model catalog" unless entries.is_a?(Array)
85
+
86
+ entries.select { |entry| entry.is_a?(Hash) }
85
87
  rescue JSON::ParserError
86
88
  raise "OpenRouter model refresh returned invalid JSON"
87
89
  end
@@ -167,17 +167,17 @@ module Kward
167
167
  }
168
168
  end
169
169
 
170
- def codex_payload(messages, tools, max_tokens: nil, model: nil, reasoning: nil)
171
- responses_payload("Codex", messages, tools, model: model, reasoning: reasoning)
170
+ def codex_payload(messages, tools, max_tokens: nil, model: nil, reasoning: nil, text_format: nil)
171
+ responses_payload("Codex", messages, tools, model: model, reasoning: reasoning, text_format: text_format)
172
172
  end
173
173
 
174
- def openai_responses_payload(messages, tools, max_tokens: nil, model: nil, reasoning: nil)
175
- payload = responses_payload("OpenAI", messages, tools, model: model, reasoning: reasoning)
174
+ def openai_responses_payload(messages, tools, max_tokens: nil, model: nil, reasoning: nil, text_format: nil)
175
+ payload = responses_payload("OpenAI", messages, tools, model: model, reasoning: reasoning, text_format: text_format)
176
176
  payload[:max_output_tokens] = max_tokens.to_i if max_tokens.to_i.positive?
177
177
  payload
178
178
  end
179
179
 
180
- def responses_payload(provider, messages, tools, model:, reasoning:)
180
+ def responses_payload(provider, messages, tools, model:, reasoning:, text_format: nil)
181
181
  parts = build_context_parts(provider, messages, tools, model: model)
182
182
  payload = {
183
183
  model: parts[:model],
@@ -194,6 +194,7 @@ module Kward
194
194
  effort = reasoning || reasoning_effort(provider)
195
195
  payload[:reasoning] = { effort: effort, summary: "auto" }
196
196
  end
197
+ payload[:text] = text_format unless text_format.nil?
197
198
  payload
198
199
  end
199
200
 
@@ -32,9 +32,11 @@ module Kward
32
32
  merge(cached_models, curated_models)
33
33
  end
34
34
 
35
- def refresh
35
+ def refresh(strict: false)
36
36
  merge(live_models, curated_models)
37
37
  rescue StandardError
38
+ raise if strict
39
+
38
40
  models
39
41
  end
40
42
 
@@ -347,7 +347,7 @@ module Kward
347
347
  when "response.content_part.added"
348
348
  codex_content_part_added(state, event["part"])
349
349
  when "response.output_text.delta", "response.refusal.delta"
350
- codex_output_text_delta(state, event["delta"], on_reasoning_delta: on_reasoning_delta, on_assistant_delta: on_assistant_delta)
350
+ codex_output_text_delta(state, event["delta"], on_assistant_delta: on_assistant_delta)
351
351
  when "response.reasoning_summary_part.added"
352
352
  codex_reasoning_summary_part_added(state, event["part"])
353
353
  when "response.reasoning_summary_text.delta"
@@ -417,7 +417,7 @@ module Kward
417
417
  state[:current_text_content_part] = item["content"].last
418
418
  end
419
419
 
420
- def codex_output_text_delta(state, delta, on_reasoning_delta: nil, on_assistant_delta: nil)
420
+ def codex_output_text_delta(state, delta, on_assistant_delta: nil)
421
421
  text = delta.to_s
422
422
  return if text.empty?
423
423
 
@@ -429,12 +429,9 @@ module Kward
429
429
  part["type"] ||= "output_text"
430
430
  text_key = part["type"] == "refusal" ? "refusal" : "text"
431
431
  part[text_key] = part[text_key].to_s + text
432
- if codex_streamable_message_item?(item)
432
+ if codex_streamable_message_item?(item) || codex_commentary_message_item?(item)
433
433
  on_assistant_delta&.call(text)
434
434
  state[:emitted_message_keys] << codex_item_key(item)
435
- elsif codex_commentary_message_item?(item)
436
- on_reasoning_delta&.call(text)
437
- state[:emitted_message_keys] << codex_item_key(item)
438
435
  end
439
436
  end
440
437
  state[:raw_content] << text
@@ -655,7 +652,7 @@ module Kward
655
652
  key = codex_item_key(item)
656
653
  if codex_commentary_message_item?(item)
657
654
  unless state[:emitted_message_keys].include?(key)
658
- on_reasoning_delta&.call(text)
655
+ on_assistant_delta&.call(text)
659
656
  state[:emitted_message_keys] << key
660
657
  end
661
658
  return
@@ -80,6 +80,8 @@
80
80
  .startup-logo { display: block; width: 7.5rem; height: 7.5rem; margin: 0 auto 1.4rem; object-fit: contain; filter: drop-shadow(0 12px 22px #0008); }
81
81
  .entry { margin: 0 0 1.15rem; animation: appear .18s ease-out; }
82
82
  @keyframes appear { from { opacity: 0; transform: translateY(3px); } }
83
+ .entry.turn-stats .label { display: none; }
84
+ .entry.turn-stats .text { color: var(--subtle); font-size: .75rem; white-space: pre-line; }
83
85
  .entry.user { display: flex; justify-content: flex-end; }
84
86
  .message { max-width: min(88%, 42rem); }
85
87
  .label { margin: 0 0 .35rem; color: var(--muted); font-size: .68rem; font-weight: 800; letter-spacing: .08em; text-transform: uppercase; }
@@ -228,7 +230,8 @@
228
230
  labelNode.textContent = label;
229
231
  const textNode = document.createElement('div');
230
232
  textNode.className = 'text';
231
- renderMarkdown(textNode, text || '');
233
+ if (role === 'turn-stats') textNode.textContent = text || '';
234
+ else renderMarkdown(textNode, text || '');
232
235
  message.append(labelNode, textNode);
233
236
  node.appendChild(message);
234
237
  transcriptInner.appendChild(node);
@@ -559,6 +562,10 @@
559
562
  events.addEventListener('answer', event => {
560
563
  if (!currentAssistant) entry('assistant', assistantLabel, JSON.parse(event.data).content);
561
564
  });
565
+ events.addEventListener('turn_summary', event => {
566
+ const data = JSON.parse(event.data);
567
+ if (data.text) entry('turn-stats', 'Turn summary', data.text);
568
+ });
562
569
  events.addEventListener('retry', event => entry('tool', 'Retry', JSON.parse(event.data).message));
563
570
  events.addEventListener('error', event => {
564
571
  if (typeof event.data === 'string' && event.data.startsWith('{')) showError(JSON.parse(event.data).message);
@@ -9,6 +9,7 @@ require "uri"
9
9
  require_relative "../agent"
10
10
  require_relative "../config_files"
11
11
  require_relative "../events"
12
+ require_relative "../transcripts/turn_summary_formatter"
12
13
  require_relative "../model/retry_message"
13
14
  require_relative "../plugins/registry"
14
15
  require_relative "../prompts/commands"
@@ -113,7 +114,12 @@ module Kward
113
114
  end
114
115
 
115
116
  def transcript_items
116
- RPC::TranscriptNormalizer.new(@conversation.messages).normalize.flat_map { |message| pan_transcript_items(message) }
117
+ RPC::TranscriptNormalizer.new(@conversation.messages).normalize.flat_map do |message|
118
+ items = pan_transcript_items(message)
119
+ text = turn_summary_text(message[:turnSummary])
120
+ items << { role: "turn-stats", label: "Turn summary", text: text } unless text.empty?
121
+ items
122
+ end
117
123
  end
118
124
 
119
125
  private
@@ -200,6 +206,8 @@ module Kward
200
206
  broadcast("turn_started", queue_payload)
201
207
  @agent.ask(prompt) do |event|
202
208
  case event
209
+ when Events::TurnSummary
210
+ broadcast("turn_summary", { summary: event.summary, text: turn_summary_text(event.summary) })
203
211
  when Events::ReasoningDelta
204
212
  broadcast("reasoning_delta", { delta: event.delta.to_s })
205
213
  when Events::ReasoningBoundary
@@ -652,6 +660,10 @@ module Kward
652
660
  { name: tool_call_name(tool_call), content: content.to_s }
653
661
  end
654
662
 
663
+ def turn_summary_text(summary)
664
+ TurnSummaryFormatter.render(summary, mode: ConfigFiles.transcript_turn_stats(@full_config))
665
+ end
666
+
655
667
  def pan_transcript_items(message)
656
668
  case message[:role] || message["role"]
657
669
  when "user"