Schemas

Configuration Models

Data structures for system configuration — profile extraction, playbook extraction, evaluation, storage, and LLM settings.

Configmodel

Root configuration object for the Emend system. Every nested type below is linked — click through to see the fields each accepts.

Prop	Type
`storage_config`required	StorageConfig
`storage_config_test`	StorageConfigTest
`agent_context_prompt`	string
`tool_can_use`	array[ToolUseConfig]
`profile_extractor_config`	ProfileExtractorConfig
`user_playbook_extractor_config`	UserPlaybookExtractorConfig
`agent_success_config`	AgentSuccessConfig
`extraction_preset`	ExtractionPreset
`window_size`	integer
`stride_size`	integer
`api_key_config`	APIKeyConfig
`llm_config`	LLMConfig
`retrieval_experiment_config`	RetrievalExperimentConfig
`retrieval_experiment_history`	array[RetrievalExperimentRecord]
`skip_should_run_check`	boolean
`enable_document_expansion`	boolean
`retrieval_floor`	RetrievalFloorConfig
`playbook_optimizer_config`	PlaybookOptimizerConfig
`lineage_gc`	LineageGCConfig
`expiry_reclamation`	ExpiryReclamationConfig
`governance_retention`	GovernanceRetentionConfig
`pending_tool_call_config`	PendingToolCallConfig
`shadow_mode_enabled`	boolean
`eval_sample_n_per_stratum`	integer
`eval_concurrency_limit`	integer
`shadow_comparison_judge_prompt_version`	string

Enterprise Offline Tuner Configuration

Enterprise get_config responses include a separate offline-tuner overlay:

{  "offline_tuner_config": {    "enabled": false  }}

Warning

offline_tuner_config is a response overlay, not a persisted shared Config field. The Python client's set_config method strips response overlays from Config model instances, including objects returned by get_config; a dictionary containing an overlay is rejected by strict shared-Config validation. The current release returns { "enabled": false } as availability status. Attempts to set enabled: true are unavailable: managed deployments return tuner_not_composed or config_reset_incomplete, and self-hosted deployments return deployment_unsupported. There is no launched or reachable offline-tuner generator, optimizer, automatic application, publication, or successor workflow.

StorageConfigmodel

StorageConfig is a union type — pass whichever variant matches your deployment. None is also accepted (no storage configured yet).

StorageConfigSQLitemodel

Local SQLite storage. The default for Local OSS.

Prop	Type
`db_path`	string

StorageConfigSupabasemodel

Supabase-backed storage for Hosted Enterprise configurations.

Prop	Type
`url`required	string
`key`required	string
`db_url`required	string
`read_url`	string
`read_key`	string
`schema`	string

StorageConfigPostgresmodel

Native PostgreSQL-backed storage for RDS or bring-your-own Postgres deployments.

Prop	Type
`type`	literal["postgres"]
`db_url`required	string
`schema`	string
`pool_size`	integer
`pool_acquire_timeout`	number
`read_db_url`	string
`read_pool_size`	integer
`read_pool_acquire_timeout`	number

StorageConfigManagedSupabasemodel

Read-only representation returned when Hosted Enterprise manages storage credentials.

Prop	Type
`managed_by`required	literal["platform"]
`schema_present`	boolean

ToolUseConfigmodel

Defines a tool that the agent can use.

Prop	Type
`tool_name`required	string
`tool_description`required	string

ProfileExtractorConfigmodel

Configuration for profile extraction from interactions.

Prop	Type
`extractor_name`	string
`extraction_definition_prompt`required	string
`context_prompt`	string
`tagging_definition_prompt`	string
`should_extract_profile_prompt_override`	string
`request_sources_enabled`	array[string]
`manual_trigger`	boolean
`window_size_override`	integer
`stride_size_override`	integer

UserPlaybookExtractorConfigmodel

Configuration for user-playbook extraction. Also exposed as PlaybookConfig (deprecated alias).

Prop	Type
`extractor_name`	string
`extraction_definition_prompt`required	string
`context_prompt`	string
`tagging_definition_prompt`	string
`aggregation_config`	PlaybookAggregatorConfig
`deduplication_config`	DeduplicationConfig
`request_sources_enabled`	array[string]
`window_size_override`	integer
`stride_size_override`	integer

PlaybookAggregatorConfigmodel

Configuration for playbook aggregation (clustering similar user playbooks into agent playbooks).

Prop	Type
`min_cluster_size`	integer
`reaggregation_trigger_count`	integer
`clustering_similarity`	float \| null
`direction_overlap_threshold`	float

DeduplicationConfigmodel

Controls bounded candidate generation when looking for existing playbooks to dedup against. The LLM consolidation stage makes the final duplicate decision.

Prop	Type
`search_threshold`	float
`search_top_k`	integer
`max_unified_content_chars`	integer

RetrievalExperimentConfigmodel

The one retrieval experiment currently assigning agent search traffic.

Prop	Type
`experiment_id`required	string
`holdout_percentage`required	float

RetrievalExperimentRecordmodel

Historical lifecycle metadata for a retrieval experiment. It contains all RetrievalExperimentConfig fields plus:

Prop	Type
`started_at`required	integer
`ended_at`	integer

RetrievalFloorConfigmodel

Configuration for the read-path relevance floor applied on unified search (/api/search). Retrieved candidates are reranked by a cross-encoder, and any result whose relevance score falls below the per-arm floor is dropped. Because filtering happens after retrieval, a search arm can return fewer than top_k results — including zero — when nothing clears its floor.

Floors are raw, model-specific cross-encoder logits, not probabilities, so they are not bounded to [0, 1] and can be negative. Leave a floor null to use the calibrated default for the reranker discovered from the private inference service; a numeric value, including 0.0, is preserved exactly.

Prop	Type
`enabled`	boolean
`pool_size`	integer
`profile_floor`	float \| null
`user_playbook_floor`	float \| null
`agent_playbook_floor`	float \| null

PlaybookOptimizerConfigmodel

Opt-in (disabled by default) GEPA-backed optimizer for playbook content. Requires exactly one assistant backend — webhook_url or assistant_script_path, not both.

Prop	Type
`enabled`	boolean
`optimize_agent_playbooks`	boolean
`optimize_user_playbooks`	boolean
`auto_update_pending_agent_playbooks`	boolean
`auto_update_user_playbooks`	boolean
`max_metric_calls`	integer
`max_turns`	integer
`early_stop_score`	float
`reflection_minibatch_size`	integer
`max_validation_windows`	integer
`min_commit_windows`	integer
`min_commit_score`	float
`min_commit_likert`	integer
`use_merge`	boolean
`max_merge_invocations`	integer
`reflection_model`	string
`webhook_url`	string
`webhook_auth_header`	string
`webhook_timeout_seconds`	integer
`webhook_max_retries`	integer
`webhook_backoff_base_seconds`	float
`assistant_script_path`	string
`assistant_script_args`	array[string]
`scheduler_jitter_seconds`	float
`cooldown_after_aborts_seconds`	integer
`abort_cooldown_threshold`	integer

PendingToolCallConfigmodel

Configuration for non-blocking pending tool calls (async information tools) during classic extraction.

Prop	Type
`enabled`	boolean
`max_pending_followups_per_scope`	integer
`pending_ttl_seconds`	integer
`dedup_cache_seconds`	integer
`prior_answer_valid_seconds`	integer
`similarity_threshold`	float
`resume_poll_interval_seconds`	float
`resume_claim_ttl_seconds`	integer
`max_resume_attempts`	integer
`max_finalization_attempts`	integer
`hmac_secrets`	array[string]
`tool_overrides`	dict[str, PendingToolCallToolOverrideConfig]

LineageGCConfigmodel

Controls reclamation of retired profile and playbook tombstones after the audit and rollback window.

Prop	Type
`enabled`	boolean
`tombstone_grace_window_days`	integer
`poll_interval_seconds`	integer

ExpiryReclamationConfigmodel

Prop	Type
`enabled`	boolean

GovernanceRetentionConfigmodel

Hosted Enterprise retention policy for audit events. These fields are accepted but do not perform audit-event reclamation in Local OSS.

Prop	Type
`audit_events_retention_enabled`	boolean
`audit_events_retention_days`	integer
`audit_events_delete_batch_limit`	integer

AgentSuccessConfigmodel

Configuration for agent success evaluation.

Prop	Type
`evaluation_name`	string
`success_definition_prompt`required	string
`tagging_definition_prompt`	string
`request_sources_enabled`	list[string]
`sampling_rate`	float
`evaluation_only_sampling_rate`	float \| null
`retrieved_learning_sampling_rate`	float \| null
`window_size_override`	integer
`stride_size_override`	integer

ExtractionPresetenum

Named presets that bundle window_size and stride_size. Accepts a string.

quick_chat — short conversations (support bots, quick Q&A). window_size=5, stride_size=3.
standard — general-purpose conversational agents. window_size=10, stride_size=5. (default)
long_form — long conversations (coding assistants, research). window_size=25, stride_size=10.
high_volume — high-traffic agents (1000+ daily interactions). window_size=15, stride_size=8.

APIKeyConfigmodel

API key configuration for LLM providers. When custom_endpoint is configured, it takes priority over all other providers for LLM completion calls (but not embeddings).

Prop	Type
`custom_endpoint`	CustomEndpointConfig
`openai`	OpenAIConfig
`anthropic`	AnthropicConfig
`openrouter`	OpenRouterConfig
`gemini`	GeminiConfig
`minimax`	MiniMaxConfig
`deepseek`	DeepSeekConfig
`dashscope`	DashScopeConfig
`zai`	ZAIConfig
`moonshot`	MoonshotConfig
`xai`	XAIConfig

CustomEndpointConfigmodel

OpenAI-compatible custom endpoint. Validated against SSRF: cloud metadata endpoints are always blocked, and when EMEND_BLOCK_PRIVATE_URLS=true, literal private IPs plus hostnames resolving to private, loopback, link-local, reserved, multicast, or unspecified IPs are blocked.

Prop	Type
`model`required	string
`api_key`required	string
`api_base`required	string

OpenAIConfigmodel

OpenAI API configuration. At least one of api_key or azure_config must be provided.

Prop	Type
`api_key`	string
`azure_config`	AzureOpenAIConfig

AzureOpenAIConfigmodel

Prop	Type
`api_key`required	string
`endpoint`required	string
`api_version`	string
`deployment_name`	string

AnthropicConfigmodel

Prop	Type
`api_key`required	string

OpenRouterConfigmodel

Prop	Type
`api_key`required	string

GeminiConfigmodel

Prop	Type
`api_key`required	string

MiniMaxConfigmodel

Prop	Type
`api_key`required	string

DeepSeekConfigmodel

Prop	Type
`api_key`required	string

DashScopeConfigmodel

Prop	Type
`api_key`required	string
`api_base`	string

ZAIConfigmodel

Prop	Type
`api_key`required	string

MoonshotConfigmodel

Prop	Type
`api_key`required	string

XAIConfigmodel

Prop	Type
`api_key`required	string

LLMConfigmodel

LLM model configuration overrides. If a field is None, the system default is used.

Prop	Type
`should_run_model_name`	string
`generation_model_name`	string
`embedding_model_name`	string
`pre_retrieval_model_name`	string

PreviousConcepts Next Enums