reduction_percent
Method
This is a method of the OptimizationResult class in the models module.Percentage of tokens reduced.
Signature
Returns
The result of the operation.
Guide
Reference
- achat
- aexecute
- agent_id
- analyze_prompt
- arun
- astart
- auto_memory
- background
- chat
- chat_with_context
- checkpoints
- clean_json_output
- clear_history
- console
- context_manager
- delete_history
- delete_history_matching
- display_name
- ephemeral
- execute
- execute_tool
- execute_tool_async
- from_template
- generate_task
- get_available_tools
- get_history_size
- get_memory_context
- get_recommended_stage
- get_rules_context
- get_skills_prompt
- handoff_to
- handoff_to_async
- iter_stream
- launch
- llm_model
- output_style
- policy
- prune_history
- query
- rag
- rag_query
- retrieval_config
- retrieve
- rules_manager
- run
- run_autonomous
- session_id
- skill_manager
- start
- store_memory
- stream_emitter
- switch_model
- thinking_budget
- add_task
- aexecute_task
- append_to_state
- arun_all_tasks
- arun_task
- astart
- clean_json_output
- clear_state
- context_manager
- current_plan
- default_completion_checker
- delete_state
- display_token_usage
- execute_task
- get_agent_details
- get_all_state
- get_all_tasks_status
- get_detailed_token_report
- get_plan_markdown
- get_state
- get_task_details
- get_task_result
- get_task_status
- get_todo_markdown
- get_token_usage_summary
- has_state
- increment_state
- launch
- restore_session_state
- run
- run_all_tasks
- run_task
- save_output_to_file
- save_session_state
- set_state
- start
- todo_list
- update_plan_step_status
- update_state
- execute
- alisten
- asay
- aspeech
- atranscribe
- console
- listen
- litellm
- say
- speech
- transcribe
- is_available
- mark_error
- mark_rate_limited
- reset
- astart
- start
- achat
- chat
- name
- rag
- run
- channel_id
- channel_type
- is_channel_allowed
- is_user_allowed
- is_webhook_mode
- command
- command_args
- is_command
- text
- channel
- content
- message_id
- sender
- bot_user
- delete_message
- edit_message
- get_agent
- get_channel
- get_user
- is_running
- on_command
- on_message
- platform
- send_message
- send_typing
- set_agent
- start
- stop
- is_bot
- user_id
- username
- fixed_total
- get_segment_budget
- history_budget
- usable
- SUPPORTED_CHUNKERS
- chunk
- chunker
- embedding_model
- aexecute
- agenerate
- console
- execute
- execute_code
- explain
- fix
- generate
- generate_code
- litellm
- refactor
- review
- aanalyze_codebase
- acreate_implementation_blueprint
- agenerate_prp
- analyze_codebase
- analyze_codebase_with_gitingest
- analyze_integration_points
- analyze_test_patterns
- build_implementation_blueprint
- compile_context_documentation
- create_implementation_blueprint
- create_quality_gates
- create_validation_framework
- execute_prp
- extract_implementation_patterns
- generate_comprehensive_prp
- generate_feature_prp
- generate_prp
- get_agent_interaction_summary
- log_debug
- perform_ast_analysis
- save_comprehensive_session_report
- save_markdown_output
- setup_logging
- setup_output_directories
- start
- for_recipe
- add_segment
- get_segment
- set_segment
- total
- capture_llm_boundary
- emergency_truncate
- estimate_tokens
- get_history
- get_last_snapshot_hook
- get_resolved_config
- get_stats
- get_tool_budget
- process
- register_snapshot_callback
- reset
- set_tool_budget
- truncate_tool_output
- optimize
- format_for_prompt
- has_citations
- from_config
- aresearch
- async_openai_client
- clarify
- follow_up
- gemini_client
- openai_client
- research
- rewrite_query
- get_all_sources
- get_citation_text
- aembed
- aembed_batch
- asimilarity
- console
- embed
- embed_batch
- find_most_similar
- litellm
- similarity
- add_profile
- get_next_profile
- get_profile
- get_retry_delay
- list_profiles
- mark_failure
- mark_success
- on_failover
- remove_profile
- reset_all
- status
- get_next_profile
- mark_failure
- mark_success
- display
- start
- stop
- client_id
- close
- connected_at
- is_connected
- receive
- send
- http_url
- is_secure
- ws_url
- broadcast
- close_session
- create_session
- emit
- get_agent
- get_session
- health
- host
- is_running
- list_agents
- list_sessions
- on_event
- port
- register_agent
- start
- stop
- unregister_agent
- add_message
- agent_id
- client_id
- close
- created_at
- get_messages
- get_state
- is_active
- last_activity
- session_id
- set_state
- default_tool_description
- default_tool_name
- execute_async
- execute_programmatic
- to_tool_function
- tool_description
- tool_name
- matches
- get_reason
- is_blocking
- should_stop
- clear
- disable_hook
- enable_hook
- enabled
- get_hooks
- has_hooks
- list_hooks
- on
- register
- register_command
- register_function
- set_global_timeout
- unregister
- allow
- ask
- block
- deny
- is_allowed
- is_denied
- aggregate_context
- execute
- execute_sync
- get_blocking_reason
- is_blocked
- registry
- achat
- aedit
- agenerate
- agenerate_image
- avariation
- chat
- edit
- generate
- generate_image
- litellm
- variation
- add
- chunker
- config
- delete
- delete_all
- get
- get_all
- get_corpus_stats
- history
- index
- markdown
- memory
- normalize_content
- reset
- search
- store
- update
- get_tools
- shutdown
- to_openai_tool
- call_tool
- run
- shutdown
- from_env
- merge
- build_context_for_task
- calculate_quality_metrics
- compute_quality_score
- delete_long_term
- delete_memories
- delete_memories_matching
- delete_memory
- delete_short_term
- finalize_task_output
- get_all_memories
- get_learn_context
- learn
- reset_all
- reset_entity_only
- reset_long_term
- reset_short_term
- reset_user_memory
- search
- search_entity
- search_long_term
- search_short_term
- search_user_memory
- search_with_quality
- store_entity
- store_long_term
- store_quality
- store_short_term
- store_user_memory
- execute
- execute_model_call
- execute_tool_call
- has_model_hooks
- has_tool_hooks
- run_after_model
- run_after_tool
- run_before_model
- run_before_tool
- add
- delete
- delete_all
- get
- get_all
- reset
- search
- update
- get_agent_manager
- get_agent_policy
- get_combined_stats
- get_session_cache
- prepare_handoff
- set_agent_policy
- aextract
- aread
- console
- extract
- litellm
- read
- close
- flush
- on_llm_call
- on_span_end
- on_span_start
- on_tool_call
- on_trace_end
- on_trace_start
- reduction_percent
- add_tool
- clear_session_permissions
- is_allowed
- is_empty
- list_tools
- load
- remove_tool
- save
- agent
- expand
- expand_basic
- expand_creative
- expand_detailed
- expand_structured
- add_abbreviation
- add_abbreviations
- agent
- rewrite
- rewrite_basic
- rewrite_contextual
- rewrite_hyde
- rewrite_multi_query
- rewrite_step_back
- rewrite_sub_queries
- format_answer_with_citations
- has_citations
- aconnect
- adisconnect
- asend_audio
- asend_text
- connect
- console
- disconnect
- on_audio
- on_error
- on_message
- receive_loop
- send_audio
- send_text
- generous
- minimal
- standard
- get_strategy
- get_token_budget
- should_retrieve
- to_knowledge_config
- to_rag_config
- all_queries
- primary_query
- docker
- e2b
- subprocess
- cleanup
- execute
- execute_file
- get_status
- is_available
- list_files
- read_file
- reset
- run_command
- sandbox_type
- start
- stop
- write_file
- output
- success
- permissive
- standard
- strict
- Agent
- add_knowledge
- add_memory
- chat
- clear_memory
- create_agent
- get_context
- get_state
- increment_state
- knowledge
- memory
- restore_state
- save_state
- search_knowledge
- search_memory
- send_message
- set_state
- check_and_add
- clear
- get_stats
- execute_callback
- execute_callback_sync
- initialize_memory
- store_in_memory
- estimate
- estimate_messages
- run
- acontent
- adownload
- agenerate
- alist
- aremix
- arun
- astart
- astatus
- await_completion
- console
- content
- download
- generate
- list
- remix
- run
- start
- status
- video_module
- wait_for_completion
- aanalyze
- acompare
- adescribe
- aextract_text
- analyze
- compare
- console
- describe
- extract_text
- litellm
- add_approval_requirement
- add_hook
- adisplay_error
- adisplay_generating
- adisplay_instruction
- adisplay_interaction
- adisplay_self_reflection
- adisplay_tool_call
- aembedding
- after_model
- after_tool
- before_model
- before_tool
- categorize_hooks
- clean_triple_backticks
- cleanup_telemetry_resources
- clear_approval_context
- configure_default_approvals
- console_approval_callback
- create_context_agent
- create_context_manager
- create_retrieval_config
- deduplicate_topics
- detect_url_scheme
- disable_performance_mode
- disable_telemetry
- display_error
- display_generating
- display_instruction
- display_interaction
- display_reasoning_steps
- display_self_reflection
- display_tool_call
- display_working_status
- embedding
- enable_performance_mode
- enable_telemetry
- encode_file_to_base64
- execute_callback
- execute_sync_callback
- force_shutdown_telemetry
- format_percent
- get_approval_callback
- get_default_registry
- get_dimensions
- get_hook_type
- get_multimodal_message
- get_permission_allowlist
- get_risk_level
- get_telemetry
- handoff
- keep_last_n_messages
- remove_all_tools
- remove_system_messages
- has_hook
- is_already_approved
- is_approval_required
- is_env_auto_approve
- is_middleware
- is_numeric_string
- is_path_like
- is_policy_string
- is_yaml_approved
- make_array_error
- make_preset_error
- mark_approved
- merge_config_with_overrides
- parse_policy_string
- parse_url_to_config
- process_task_context
- process_video
- prompt_with_handoff_instructions
- register_approval_callback
- register_display_callback
- remove_approval_requirement
- remove_hook
- request_approval
- require_approval
- reset_yaml_approved_tools
- resolve
- resolve_autonomy
- resolve_batch
- resolve_caching
- resolve_context
- resolve_execution
- resolve_guardrail_policies
- resolve_guardrails
- resolve_hooks
- resolve_knowledge
- resolve_memory
- resolve_output
- resolve_planning
- resolve_reflection
- resolve_routing
- resolve_skills
- resolve_web
- set_approval_callback
- set_default_registry
- set_permission_allowlist
- set_yaml_approved_tools
- suggest_similar
- trace
- track_workflow
- wrap_model_call
- wrap_tool_call
Functions
reduction_percent
Percentage of tokens reduced.
⌘I