Query memory with recall()

recall() is cognee's query API. It picks a search strategy, searches the graph and (with a session) the session cache, and returns a list of tagged results.

import cognee

results = await cognee.recall("Where was Einstein born?")
for r in results:
    print(r.source, r.text)  # e.g. "graph", "Einstein was born in Ulm."

Without datasets it searches every dataset the user can read. Pass datasets=["research"] (names) or dataset_ids=[...] (UUIDs, which win) to narrow it; that is also faster.

Use it

How the search type is picked

1. An explicit query_type=SearchType.X always wins.
2. Otherwise, with no usable LLM key, CHUNKS (plain vector search).
3. Otherwise the router (auto_route=True, the default). It is three regex rules, first match wins, no LLM call:
   • the whole query is one "quoted phrase" → CHUNKS_LEXICAL
   • mentions coding rules/standards/conventions or code-review guidelines → CODING_RULES
   • an absolute date behind a time preposition (in 2019, before 1900, on 7 November 1867) → TEMPORAL
4. Everything else → HYBRID_COMPLETION.

A routed type (never a pinned one) that the backend rejects, or a routed CHUNKS_LEXICAL / CODING_RULES that returns nothing, is retried once as HYBRID_COMPLETION. The router never picks CYPHER.

from cognee import SearchType

await cognee.recall("What changed in v2?", query_type=SearchType.GRAPH_COMPLETION)

Search types

The full list is cognee/modules/search/types/SearchType.py; the type-to-retriever table is cognee/modules/retrieval/README.md.

Scope: which sources are searched

scope is one of, or a list of: graph, session, session_first, trace, session_context, all, tools, code. all means graph + session + trace + session_context; tools and code are never included implicitly.

When scope is omitted:

Session and trace search is keyword overlap, not embeddings.

Filters and knobs

• top_k=15: per dataset, not in total. The default HYBRID caps each lane at min(top_k, 10); set chunks_top_k / entities_top_k / facts_top_k in retriever_specific_config to go higher.
• node_name=["AI"] (+ node_name_filter_operator="OR"|"AND"): restricts graph/chunk/completion types to data remembered with that node_set. SUMMARIES, CHUNKS_LEXICAL, GRAPH_REPORT, CYPHER, NATURAL_LANGUAGE, CODE and SKILLS ignore it; CODING_RULES treats it as the rules node-set name.
• system_prompt / system_prompt_path: change the answering prompt.
• response_model=MyPydanticModel: structured answer, on r.structured.
• include_references=True: attach the document chunks that support each graph edge used (needs EDGE_EVIDENCE_ENABLED=true, the default).
• only_context=True: return what the LLM would have received instead of an answer. r.text is the rendered user prompt, r.system_prompt the system prompt. Pin query_type when you use it.
• retriever_specific_config={...}: retriever-only options. For the agentic extras (skills, tools, max_iter) and node_type, call cognee.search() instead, which takes them as parameters.

Reading the results

Each item is a Pydantic model with a source discriminator: graph, session, trace, session_context, code, tools, skills, or system. Graph items carry text (always renderable), search_type, kind, score, dataset_id / dataset_name, metadata, raw, and structured. A system item is a status marker, not data (see "memory warming up" below).

When a query sounds procedural ("how do I…", "runbook", "steps to…") and exactly one dataset is targeted, recall also runs a small SKILLS lookup and appends hits with source="skills". Disable with SKILL_GATE_ENABLED=false.

Pitfalls

• Permissions change what you get back. With no datasets, recall searches only datasets the user can read, so a user without grants gets one source="system" memory_warming_up marker from a graph-only recall (with only_context=True or RECALL_WARMUP_SHORTCIRCUIT=false it gets []; with session sources included only the graph lane is empty, and session and trace hits still come back). Asking for a dataset id the user cannot read raises PermissionDeniedError (HTTP 403). Dataset names resolve only among the user's own datasets, so a name that is not theirs (even one shared with them) raises DatasetNotFoundError; use dataset_ids for shared datasets. See the cognee-permissions skill.
• "Memory warming up". On an empty graph, a graph-only recall (no session sources, not only_context) returns one source="system" item with status="memory_warming_up" (or "build_failed" plus error_message) instead of results; a multi-source recall just returns no graph results. Wait for the remember to finish, or check why it failed.
• Hybrid silently becomes graph completion when you pass neighborhood_depth or feedback_influence > 0 (including a nonzero DEFAULT_FEEDBACK_INFLUENCE), or the chunk collection is missing. node_name stays on hybrid, which filters to that node set. search() also defers for a custom node_type or node_name with node_type=None; recall() has no node_type. wide_search_top_k and triplet_distance_penalty with hybrid raise InvalidHybridSearchConfig; pin GRAPH_COMPLETION to use them.
• SKILLS and AGENTIC_COMPLETION need exactly one dataset. For SKILLS, search() raises unless exactly one dataset is given. recall() runs SKILLS per dataset with access control on (zero datasets gives [], not an error) and raises only with access control off and not exactly one dataset. For AGENTIC_COMPLETION only search() checks it up front, so call it through search() with one dataset.
• code_query without scope="code" raises, and scope="tools" also needs TOOL_CALLS_ENABLED=true.
• Latency. Most completion types make one LLM call; COT, DECOMPOSITION, CONTEXT_EXTENSION, GRAPH_SUMMARY_COMPLETION, TEMPORAL, FEELING_LUCKY (one call to pick the type, then the chosen type's) and AGENTIC_COMPLETION (a loop of up to max_iter, default 6) make more. NATURAL_LANGUAGE makes one (no answer call) and retries only on an empty or failed query, up to 3 attempts. With CACHING and AUTO_FEEDBACK on (defaults), each answered turn adds one analysis call, even without a session_id. Set AUTO_FEEDBACK=false for low-latency reads (see the cognee-performance skill).

recall() or search()?

Use recall(). Drop to cognee.search() only for agentic parameters (skills, tools, max_iter, node_type) as first-class arguments, raw SearchResult objects, or a pinned type with no router. search() never searches the session cache; its session_id only adds conversation history to the prompt. Full guide: docs/recall-vs-search.md.

How it works

recall() resolves scope and search type, then calls the same authorized search search() uses: datasets resolve through the permission layer with read, one search per dataset runs concurrently, and results are normalized and tagged.

• Entry point, scope and type resolution: cognee/api/v1/recall/recall.py
• Router: cognee/api/v1/recall/query_router.py
• Skill gate: cognee/api/v1/recall/skill_gate.py
• Result types: cognee/modules/recall/types/RecallResponse.py, SearchResultItem.py
• Scope names: cognee/memory/entries.py:normalize_scope
• Warm-up config (RECALL_WARMUP_*): cognee/modules/recall/config.py
• Core search and fan-out: cognee/modules/search/methods/search.py
• Hybrid fallback rules: cognee/modules/search/methods/hybrid_deferral.py
• Registry: cognee/modules/search/methods/get_search_type_retriever_instance.py

Examples in examples/guides/: recall_core.py, hybrid_retrieval_recall.py, references_example.py, temporal_recall.py, sessions.py.

Extending it

Adding a search type, per cognee/modules/retrieval/README.md:

1. Write the retriever in cognee/modules/retrieval/ (subclass BaseRetriever or a completion base).
2. Add the SearchType member and its search_core_registry entry.
3. Add a row to the README table. A unit test (cognee/tests/unit/modules/retrieval/retriever_readme_index_test.py) fails if the table and registry disagree.
4. Optional: add it to SEARCH_TYPE_CHOICES in cognee/cli/config.py for the CLI, or a regex rule to query_router.py for auto-routing. Never route a type that can write.