Agent Hints
Configuration directives that guide the extraction and retrieval models toward domain-specific behavior. Agent hints are set in the MACA configuration and can include custom terminology, extraction priorities, and retrieval preferences. For example, a medical support agent might have hints to prioritize extracting medication names and dosages.Bootstrap Key
A one-time credential used to establish the initial mTLS connection between the SDK and Synap Cloud. Bootstrap keys are generated via the Dashboard or Instances API, are short-lived, and can only be used once.Client
The top-level organizational entity in Synap, representing your company or team. A client owns one or more instances and is identified by a client ID in the formatcli_<hex16>. All API keys, billing, and team management are scoped to the client level. See Clients and Instances.
Compaction
The process of condensing a conversation’s accumulated context into a shorter summary while preserving the most important information. Compaction reduces token usage when injecting context into LLM prompts. Strategies includesummarize, extract, hierarchical, and temporal. See Context Compaction and the Context API.
Confidence Score
A numerical value between0.0 and 1.0 assigned to each extracted memory indicating how certain the extraction model is that the information is accurate and correctly categorized. The minimum confidence threshold is configurable in MACA (default: 0.7). Memories below the threshold are discarded during ingestion.
Context Budget
The maximum number of tokens allowed in a context response. The context budget prevents the retrieval system from returning more data than the downstream LLM can process. Configurable per instance in MACA (default:4096 tokens).
Context Response
The structured output returned by the Context Fetch endpoint. Contains ranked memories organized by type (facts, preferences, episodes, emotions), along with metadata about query performance and token usage.Correlation ID
A unique request identifier included in every API response via theX-Correlation-Id header. The correlation ID traces a request through the entire Synap processing pipeline, including asynchronous jobs. Always include the correlation ID when contacting support. Format: req_<uuid>.
Customer
A scoping entity within a client that represents an end-customer or account. Memories scoped to a customer are accessible by all users within that customer but isolated from other customers. Part of the scope chain. See Scopes.Entity Resolution
The process of identifying and linking mentions of the same real-world entity across different conversations and documents. For example, “John”, “John Smith”, and “my manager” may all resolve to the same person entity. Synap uses semantic matching to resolve entities. Unresolved entities are auto-registered at the CUSTOMER scope. See Entities and Resolution.Episode
A memory type representing a narrative account of an event or interaction. Episodes capture “what happened” with temporal context. Example: “User described a frustrating experience trying to set up their printer last weekend.” See Memory Lifecycle.Emotion
A memory type representing an emotional state, sentiment, or feeling expressed in a conversation. Example: “User expressed excitement about their upcoming vacation.” See Memory Lifecycle.Fact
A memory type representing objective, factual information extracted from conversations. Facts are statements about the world, the user, or their circumstances. Example: “User works at Acme Corp as a senior engineer.” See Memory Lifecycle.Graph Store
One of two storage backends used by Synap. The graph store maintains entity relationships and traversal paths, enabling queries like “find all memories related to this person’s workplace.” Graph queries complement vector store similarity search for deeper contextual retrieval. See Storage Engines.Ingestion
The asynchronous pipeline that processes raw documents into structured memories. The pipeline has four stages: extraction (identifying memory-worthy content), categorization (classifying into memory types), entity resolution (linking entity mentions), and storage (writing to vector and graph stores). See Memory Lifecycle and the Ingestion SDK.Instance
An isolated memory store for a single AI agent, belonging to a client. Each instance has its own MACA configuration, storage namespaces, authentication credentials, and usage metrics. Instances are identified byinst_<hex16>. See Clients and Instances and the Instances API.
MACA (Memory Architecture Configuration)
The configuration system that controls how an instance processes, stores, and retrieves memories. MACA encompasses storage settings (vector/graph namespaces, retention), ingestion rules (categories, extraction, chunking), and retrieval behavior (modes, ranking, budget). Configuration changes go through a review workflow. See Memory Architecture and the Configuration API.Memory
A unit of structured knowledge extracted from a document during ingestion. Memories are typed as facts, preferences, episodes, emotions, or temporal events. Each memory has a confidence score, scope, and linked entities. See Memory Lifecycle.Memory Category
The classification of a memory into one of the supported types:fact, preference, episode, emotion, or temporal_event. Categories are determined during the categorization stage of ingestion. The set of active categories is configurable in MACA.
mTLS (Mutual TLS)
The authentication mechanism used for gRPC streaming connections between the SDK and Synap Cloud. Both the client (SDK) and server (Cloud) present certificates to verify each other’s identity. Certificates are obtained through the bootstrap key flow and are valid for 7 days with a 48-hour grace period. See Authentication.Namespace
An isolated storage partition within the vector store or graph store. Each instance gets its own namespace to ensure memory isolation. Namespace identifiers are auto-generated based on the instance ID.Preference
A memory type representing a user’s opinions, likes, dislikes, or behavioral preferences. Example: “Prefers light roast coffee, especially pour-over method.” See Memory Lifecycle.Primary Scope
The most specific scope in the scope chain that a memory belongs to. Determines the default visibility of the memory. A memory withuser_id set has USER as its primary scope; with only customer_id, it has CUSTOMER scope.
Ranking Signals
The factors used to order retrieved memories by relevance during context fetch. Available signals includerecency (newer memories ranked higher), relevance (semantic similarity to the query), and confidence (higher confidence memories ranked higher). Configurable in MACA.
Retention Policy
Rules governing how long memories are stored and when they are evicted. Configured per instance in MACA. Options includedefault_ttl_days (time-to-live) and max_memories_per_scope (capacity limit). Setting TTL to null means indefinite retention.
Review Queue
The workflow system for approving MACA configuration changes. When a configuration is submitted or updated, it enters the review queue with statuspending_review. A reviewer can approve, reject, or request_changes. Only approved configurations can be applied.
Scope
A visibility boundary that determines who can access a memory. Synap supports four scope levels: USER (single user), CUSTOMER (all users in a customer account), CLIENT (all users across the organization), and WORLD (global). See Scopes.Scope Chain
The hierarchical ordering of scopes from narrowest to broadest: USER > CUSTOMER > CLIENT > WORLD. During context fetch, Synap searches from the narrowest applicable scope outward, prioritizing more specific memories. See Scopes.Temporal Event
A memory type representing a time-bound event with specific dates, deadlines, or scheduled occurrences. Example: “User has a dentist appointment on January 20th at 2pm.” See Memory Lifecycle.Vector Store
One of two storage backends used by Synap. The vector store indexes memory embeddings for fast similarity search. See Storage Engines.Webhook
An HTTP callback that delivers real-time event notifications to your application. Synap supports five event types:conversation.started, conversation.ended, context.retrieved, config.applied, and compaction.completed. Webhook payloads are signed with HMAC-SHA256 for verification. See Webhooks.