diff --git a/README.md b/README.md
index e839366f..7610f065 100644
--- a/README.md
+++ b/README.md
@@ -138,6 +138,7 @@ agent = Agent(model="oci:openai.gpt-5.5", tools=[retriever.as_tool()])
 
 → [OCI GenAI models](https://locusagents.ai/how-to/oci-models/) ·
 [Dedicated AI Clusters](https://locusagents.ai/how-to/oci-dac/) ·
+[Oracle Database 26ai](https://locusagents.ai/concepts/oracle-26ai/) ·
 [Notebook 01 — OCI transports](https://locusagents.ai/notebooks/notebook_01_oci_transports/) ·
 [Notebook 38 — Oracle 26ai RAG](https://locusagents.ai/notebooks/notebook_38_oracle_26ai_rag/) ·
 [Notebook 42 — Oracle 26ai checkpointer](https://locusagents.ai/notebooks/notebook_42_oracle_26ai_checkpointer/) ·
@@ -150,7 +151,7 @@ agent = Agent(model="oci:openai.gpt-5.5", tools=[retriever.as_tool()])
 | | |
 |---|---|
 | **[Tools and MCP](https://locusagents.ai/concepts/tools/)** | `@tool` on any function; arguments coerced to the declared types whatever the model sends; `@tool(idempotent=True)` dedupes on `(name, args)`; `MCPClient` consumes MCP servers and `LocusMCPServer` exposes yours. |
-| **[Memory](https://locusagents.ai/concepts/checkpointers/)** | 10 checkpoint backends (Oracle Database 26ai and its versioned saver, OCI Object Storage, PostgreSQL, MySQL, Redis, OpenSearch, S3-compatible, in-memory, file, HTTP); `OracleStore`, Oracle agent memory and the portable `LLMMemoryManager` for long-term memory. |
+| **[Memory](https://locusagents.ai/concepts/memory/)** | 10 checkpoint backends (Oracle Database 26ai and its versioned saver, OCI Object Storage, PostgreSQL, MySQL, Redis, OpenSearch, S3-compatible, in-memory, file, HTTP); `OracleStore`, Oracle agent memory and the portable `LLMMemoryManager` for long-term memory. |
 | **[RAG](https://locusagents.ai/concepts/rag/)** | Six vector stores (Oracle 26ai, pgvector, Qdrant, Chroma, OpenSearch, in-memory); OCI, in-database, OpenAI and Cohere embeddings; OCI Cohere, local and Cohere rerankers; `retriever.as_tool()`. |
 | **[Grounded reasoning](https://locusagents.ai/concepts/gsar/)** | GSAR partitions every claim as grounded, ungrounded, contradicted or complementary against typed evidence; below threshold the agent regenerates, replans or abstains. Reflexion scores every turn. |
 | **[Guardrails and the admission gate](https://locusagents.ai/concepts/control-layer/)** | Input, output and tool-call guardrails, steering (LLM-as-judge), trace hygiene, and `admit()` / `approve()`: allow, hold for a human, or deny, on a hash-chained audit trail. |
diff --git a/docs/api/models.md b/docs/api/models.md
index ae695cd7..4022fcbc 100644
--- a/docs/api/models.md
+++ b/docs/api/models.md
@@ -14,7 +14,7 @@ For the concepts, start with [Model providers](../concepts/models.md).
 ## Registry
 
 String factory — routes `"openai:gpt-4o"`,
-`"anthropic:claude-sonnet-4-6"`, etc. to the right client.
+`"anthropic:claude-sonnet-5-5"`, etc. to the right client.
 
 ::: locus.models.registry.get_model
 ::: locus.models.registry.list_providers
diff --git a/docs/capabilities.md b/docs/capabilities.md
index 73aa1673..208eb8b1 100644
--- a/docs/capabilities.md
+++ b/docs/capabilities.md
@@ -274,8 +274,8 @@ await ctx.actions.execute(
 |---|---|---|
 | OCI Generative AI | OpenAI, Meta Llama, xAI Grok, Mistral, Google Gemini and Cohere models served on OCI (on-demand or Dedicated AI Cluster) | `locus.models.providers.oci` · [OCI Generative AI](concepts/providers/oci.md) |
 | OpenAI | All commercial models (gpt-5.5, o-series, etc) | `locus.models.native.openai` · [OpenAI](concepts/providers/openai.md) |
-| Anthropic | Claude 4.x (e.g. `claude-sonnet-4-6`) — direct API | `locus.models.native.anthropic` · [Anthropic](concepts/providers/anthropic.md) |
-| Auto-routing | `get_model("anthropic:claude-sonnet-4-6")` picks transport from id | `locus.models.registry.get_model` |
+| Anthropic | Claude 5.x (e.g. `claude-sonnet-5-5`) — direct API | `locus.models.native.anthropic` · [Anthropic](concepts/providers/anthropic.md) |
+| Auto-routing | `get_model("anthropic:claude-sonnet-5-5")` picks transport from id | `locus.models.registry.get_model` |
 | Resilience helpers | Failover classification, credential pooling, prompt-cache breakpoints, rate-limit buckets | `locus.models.failover` · `locus.models.pooled` · `locus.models.caching` · `locus.models.rate_limits` · [Resilience](concepts/providers/resilience.md) |
 
 ## Skills + Playbooks
diff --git a/docs/concepts/checkpointers.md b/docs/concepts/checkpointers.md
index 8bf61a32..e1a88467 100644
--- a/docs/concepts/checkpointers.md
+++ b/docs/concepts/checkpointers.md
@@ -81,6 +81,32 @@ lifecycle policies handle retention, IAM handles auth.
 
 ## Getting started
 
+### Oracle Database 26ai: `oracle_checkpointer()`
+
+Durable threads in the same database as the data the agent acts on. The
+factory returns a ready checkpointer; `OracleCheckpointSaver` is the
+versioned alternative that keeps every checkpoint of a thread (see
+[Oracle Database 26ai](oracle-26ai.md)).
+
+```python
+import os
+
+from locus.agent import Agent
+from locus.memory.backends import oracle_checkpointer
+
+cp = oracle_checkpointer(
+    dsn=os.environ["ORACLE_DSN"],
+    user=os.environ["ORACLE_USER"],
+    password=os.environ["ORACLE_PASSWORD"],
+    wallet_location=os.environ["ORACLE_WALLET"],        # Autonomous Database wallet
+    wallet_password=os.environ.get("ORACLE_WALLET_PASSWORD", ""),
+    table_name="locus_checkpoints",
+)
+agent = Agent(model="oci:openai.gpt-5.5", checkpointer=cp)
+agent.run_sync("My name is Alice.", thread_id="t1")
+agent.run_sync("What's my name?", thread_id="t1")   # resumes from the database
+```
+
 ### Local: `FileCheckpointer`
 
 ```python
diff --git a/docs/concepts/hooks.md b/docs/concepts/hooks.md
index 4adb9d28..63203979 100644
--- a/docs/concepts/hooks.md
+++ b/docs/concepts/hooks.md
@@ -180,7 +180,7 @@ from locus.models import AnthropicModel
 
 agent = Agent(
     ...,
-    hooks=[SteeringHook(model=AnthropicModel(model="claude-sonnet-4-6"))],
+    hooks=[SteeringHook(model=AnthropicModel(model="claude-sonnet-5-5"))],
 )
 ```
 
diff --git a/docs/concepts/memory-manager.md b/docs/concepts/memory-manager.md
index 0be62ebc..4b08cdb7 100644
--- a/docs/concepts/memory-manager.md
+++ b/docs/concepts/memory-manager.md
@@ -92,7 +92,7 @@ async def main():
     )
 
     # Session 1 — agent learns a standing rule
-    async for event in agent.run("Refunds over $500 need manager approval — page the duty manager first."):
+    async for event in agent.run("Never refund over $500 without manager approval — page the duty manager first."):
         ...
 
     # Session 2 — the agent already knows
@@ -172,9 +172,9 @@ The SDK exposes memory along a spectrum, from "works everywhere" to
 
 | Path | Manager class | When to pick it |
 |---|---|---|
+| Oracle-native | `OracleAgentMemoryManager` ([`oracleagentmemory`](https://pypi.org/project/oracleagentmemory/)) | You run on Oracle Database 26ai and want Oracle's official agent-memory client — context-card summaries, scoped retrieval by user / agent / thread, LongMemEval-benchmarked extraction |
 | Portable / multi-backend | `LLMMemoryManager` + any `BaseStore` (`InMemoryStore`, `HolographicStore`, `PgMemory`, or a custom subclass) | You need backend portability, an LLM-free extractor, or a test-friendly path |
 | Managed | `Mem0MemoryManager` ([`mem0ai`](https://pypi.org/project/mem0ai/)) | You want a managed memory layer with LLM-tuned recall and scoped retrieval by `user_id` without writing your own extractor |
-| Oracle-native | `OracleAgentMemoryManager` ([`oracleagentmemory`](https://pypi.org/project/oracleagentmemory/)) | You run on Oracle Database 26ai and want Oracle's official agent-memory client — context-card summaries, scoped retrieval by user / agent / thread, LongMemEval-benchmarked extraction |
 
 ### Managed memory — `Mem0MemoryManager`
 
@@ -205,6 +205,36 @@ pip install 'locus-sdk[mem0]'
 When you need a fully-deterministic LLM-free extractor or a custom
 backend, fall back to the portable path below.
 
+### Oracle Database 26ai: `OracleStore` and `OracleAgentMemoryManager`
+
+Two ways to keep long-term memory in Oracle Database 26ai. `OracleStore` is a
+`BaseStore`, so it slots under `LLMMemoryManager` with the portable
+extractor and per-user scoping; `OracleAgentMemoryManager` hands the whole
+job to Oracle's own `oracleagentmemory` client (extraction, context cards,
+retrieval scoped by user, agent and thread, all inside the database).
+
+```python
+import os
+
+from locus.agent import Agent
+from locus.memory import LLMMemoryManager
+from locus.memory.store_backends import OracleStore
+
+store = OracleStore(
+    dsn=os.environ["ORACLE_DSN"],
+    user=os.environ["ORACLE_USER"],
+    password=os.environ["ORACLE_PASSWORD"],
+    wallet_location=os.environ["ORACLE_WALLET"],
+    wallet_password=os.environ.get("ORACLE_WALLET_PASSWORD", ""),
+)
+agent = Agent(model="oci:openai.gpt-5.5", memory_manager=LLMMemoryManager(store=store))
+```
+
+Notebooks [37](../notebooks/notebook_37_oracle_store.md) and
+[44](../notebooks/notebook_44_oracle_agent_memory.md) run both against a
+live database; [Oracle Database 26ai](oracle-26ai.md) lists the rest of the
+stack.
+
 ### Portable path (any BaseStore backend)
 
 `LLMMemoryManager` works against any `BaseStore` implementation. Use
diff --git a/docs/concepts/memory.md b/docs/concepts/memory.md
new file mode 100644
index 00000000..4850c304
--- /dev/null
+++ b/docs/concepts/memory.md
@@ -0,0 +1,118 @@
+---
+title: Memory — short-term and long-term
+description: "The three memories of a Locus agent — working memory, thread memory and long-term memory — plus knowledge retrieval, and how Oracle Database 26ai backs each of them."
+---
+
+# Memory: short-term and long-term
+
+An agent remembers at three different time scales, and Locus keeps them as
+three separate surfaces so you can choose a backend for each. Oracle
+Database 26ai can back all of them, in the same database as the business
+data the agent acts on.
+
+| Layer | What it holds | Lifetime | Locus surface | Oracle Database 26ai |
+|---|---|---|---|---|
+| **Working memory** | The messages of the current run: prompt, tool calls, observations | One run | `Agent(conversation_manager=…)`: `SlidingWindowManager`, `SummarizingManager`, `ContextCompactor` | In-process, nothing to persist |
+| **Short-term memory** (the thread) | The full state of one conversation: messages, tool results, iteration, so a thread can resume in another process or after a crash | The thread | `Agent(checkpointer=…)` and `thread_id=` on every call | `oracle_checkpointer()` for durable threads; `OracleCheckpointSaver` keeps every version for history and resume-from-any-point |
+| **Long-term memory** | Facts that outlive any thread: preferences, standing rules, what was learned handling earlier cases | Until you delete it | `Agent(memory_manager=…)`: `LLMMemoryManager(store=…)` or `OracleAgentMemoryManager` | `OracleStore` as the store; `OracleAgentMemoryManager` over Oracle's `oracleagentmemory` client |
+| **Knowledge** (RAG) | Documents the agent retrieves rather than remembers | Your corpus | `RAGRetriever(...).as_tool()` | `OracleVectorStore`, `OracleInDBChunker`, `OracleInDBEmbeddings`, `OracleADBLoader` |
+
+## Working memory
+
+Working memory is the context window. It is not stored anywhere; the
+question is what to drop when a long run no longer fits. A
+`ConversationManager` decides: `SlidingWindowManager` keeps the last *n*
+messages, `SummarizingManager` folds older turns into a summary, and
+`ContextCompactor` archives tool output and replaces it with a reference.
+See [the agent loop](agent-loop.md) for where it runs.
+
+```python
+from locus.agent import Agent
+from locus.memory.conversation import SlidingWindowManager
+
+agent = Agent(model="oci:openai.gpt-5.5", conversation_manager=SlidingWindowManager(window_size=20))
+```
+
+## Short-term memory: the thread
+
+A checkpointer saves the agent's state under a `thread_id` and loads it back
+on the next call, from any process. That is what makes a conversation
+survive a restart, a deploy, or a hand-off to the next shift.
+
+```python
+import os
+
+from locus.agent import Agent
+from locus.memory.backends import oracle_checkpointer
+
+cp = oracle_checkpointer(
+    dsn=os.environ["ORACLE_DSN"],
+    user=os.environ["ORACLE_USER"],
+    password=os.environ["ORACLE_PASSWORD"],
+    wallet_location=os.environ["ORACLE_WALLET"],
+    wallet_password=os.environ.get("ORACLE_WALLET_PASSWORD", ""),
+    table_name="support_threads",
+)
+agent = Agent(model="oci:openai.gpt-5.5", checkpointer=cp)
+agent.run_sync("My name is Alice.", thread_id="case-100")
+agent.run_sync("What is my name?", thread_id="case-100")  # another process, same thread
+```
+
+`oracle_checkpointer()` keeps the latest state of each thread.
+`OracleCheckpointSaver` keeps every version, so you can list a thread's
+history and resume from any earlier checkpoint. Ten backends ship in total;
+[Checkpointers](checkpointers.md) compares them.
+
+## Long-term memory
+
+Long-term memory is what the agent keeps *across* threads: Alice's
+preferences should be known in her next case, not only in this one. A
+`MemoryManager` runs two hooks around every run: before the first model
+call it retrieves relevant memories into the system prompt; after the run
+it extracts durable facts and persists them.
+
+```python
+import os
+
+from locus.agent import Agent
+from locus.memory import LLMMemoryManager
+from locus.memory.store_backends import OracleStore
+
+store = OracleStore(
+    dsn=os.environ["ORACLE_DSN"],
+    user=os.environ["ORACLE_USER"],
+    password=os.environ["ORACLE_PASSWORD"],
+    wallet_location=os.environ["ORACLE_WALLET"],
+    wallet_password=os.environ.get("ORACLE_WALLET_PASSWORD", ""),
+)
+agent = Agent(model="oci:openai.gpt-5.5", memory_manager=LLMMemoryManager(store=store))
+
+agent.run_sync("I'm Alice. I always want quotes in EUR, never in USD.", thread_id="case-100")
+agent.run_sync("Which currency do I want my quotes in?", thread_id="case-200")  # new thread: EUR
+```
+
+Two managers ship. `LLMMemoryManager` works over any `BaseStore`, with
+Oracle Database 26ai through `OracleStore`, and lets you scope memories per
+user or tenant; out of the box it extracts with a small heuristic (standing
+rules such as "always … never …", who the user is, project context, links),
+and you pass `extract_fn` to let a model decide what is worth keeping. `OracleAgentMemoryManager` wraps Oracle's own agent-memory
+client (`oracleagentmemory`): extraction, context-card summaries and
+retrieval scoped by user, agent and thread, all inside the database.
+[Long-term memory manager](memory-manager.md) covers both.
+
+## Knowledge is not memory
+
+What the agent should *look up* belongs in retrieval, not in memory: policy
+pages, runbooks, past resolutions. [RAG](rag.md) covers embedders, vector
+stores and `retriever.as_tool()`; on Oracle Database 26ai the whole pipeline
+can run inside the database.
+
+## Why one database
+
+Threads, memories and vectors can each live in a different system; on
+Oracle Database 26ai they live in one, next to the tables the agent reads
+and writes through its tools. That means one set of credentials, one backup
+and retention policy, one place to audit what the agent knew when it acted,
+and no copy of customer data in a separate memory service.
+[Oracle Database 26ai](oracle-26ai.md) lists every primitive;
+notebooks 37 to 44 run each one.
diff --git a/docs/concepts/models.md b/docs/concepts/models.md
index 89b6134e..5e83a3a3 100644
--- a/docs/concepts/models.md
+++ b/docs/concepts/models.md
@@ -8,9 +8,9 @@ OpenAI-compatible prefixes such as `ollama:`, `vllm:`, `groq:` or
 
 ```python
 # tools, system_prompt, and other kwargs are the same across providers
-Agent(model="oci:openai.gpt-5", tools=[lookup_order, issue_refund])             # OCI Generative AI
+Agent(model="oci:openai.gpt-5.5", tools=[lookup_order, issue_refund])           # OCI Generative AI
 Agent(model="openai:gpt-4o", tools=[lookup_order, issue_refund])                # OpenAI direct
-Agent(model="anthropic:claude-sonnet-4-6", tools=[lookup_order, issue_refund])  # Anthropic direct
+Agent(model="anthropic:claude-sonnet-5-5", tools=[lookup_order, issue_refund])  # Anthropic direct
 Agent(model="openrouter:provider/model-id", tools=[lookup_order, issue_refund]) # OpenRouter
 Agent(model="together:organization/model-id", tools=[lookup_order, issue_refund]) # Together.ai
 ```
@@ -56,10 +56,11 @@ locus.models
 │
 ├── oci:                                   ── OCI Generative AI · auto-routed transport
 │   ├─ V1 (default)            — OCIChatCompletionsModel on /openai/v1:
-│   │                           OpenAI · Meta · xAI · Mistral · Gemini ·
-│   │                           Cohere (non-R) — SSE streaming, day-0 models
-│   ├─ SDK                     — OCIModel for cohere.command-r* and
-│   │                           Dedicated AI Cluster endpoint OCIDs
+│   │                           OpenAI · Meta · xAI · Google — SSE
+│   │                           streaming, new models as OCI lists them
+│   ├─ SDK                     — OCIModel for every cohere.* id (Command R
+│   │                           and Command A; OCI's /openai/v1 rejects
+│   │                           them) and Dedicated AI Cluster endpoint OCIDs
 │   └─ Responses (opt-in)      — OCIResponsesModel, server-stateful
 │
 ├── openai:                                ── OpenAI direct · OpenAIModel
@@ -68,12 +69,12 @@ locus.models
 │
 ├── anthropic:                             ── Anthropic direct · AnthropicModel
 │   ├─ Claude family          — opus · sonnet · haiku
-│   └─ prompt caching         — opt-in (prompt_cache=True), non-streaming
-│                               calls only: marks the system prompt + tool
-│                               catalog; once that prefix passes the model's
-│                               minimum cacheable length, later turns read
-│                               it at ~1/10th input cost while the ~5-min
-│                               cache holds
+│   └─ prompt caching         — opt-in (prompt_cache=True), on complete()
+│                               and stream() alike: marks the system prompt
+│                               + tool catalog; once that prefix passes the
+│                               model's minimum cacheable length, later
+│                               turns read it at ~1/10th input cost while
+│                               the ~5-min cache holds
 │
 ├── OpenAI-compatible prefixes             ── OpenAIModel at a routed base_url
 │   ├─ self-hosted            — ollama · vllm · lmstudio · llamacpp
@@ -131,7 +132,7 @@ register_provider("audited", lambda model_id, **kw: AuditedModel(model_id, **kw)
 agent = Agent(model="audited:internal-llm", tools=[lookup_order, issue_refund])
 ```
 
-Source: [`register_provider` in `models/registry.py:21`](https://github.com/oracle/locus/blob/main/src/locus/models/registry.py#L21).
+Source: [`register_provider` in `models/registry.py`](https://github.com/oracle/locus/blob/main/src/locus/models/registry.py).
 
 ## Credential pooling & rotation
 
@@ -152,7 +153,7 @@ pool = CredentialPool([
 ])
 
 def _build(cred: Credential) -> AnthropicModel:
-    return AnthropicModel(model="claude-sonnet-4-6", api_key=cred.api_key)
+    return AnthropicModel(model="claude-sonnet-5-5", api_key=cred.api_key)
 
 agent = Agent(
     model=CredentialPoolModel(pool=pool, build_model=_build),
@@ -171,7 +172,7 @@ Source:
 ## Notebook
 
 [`notebook_11_model_providers.py`](https://github.com/oracle/locus/blob/main/examples/notebook_11_model_providers.py)
-runs the same agent — a security-operations triage example — against OpenAI and
+runs the same cloud-operations agent against OCI Generative AI, OpenAI and
 Anthropic by swapping one string.
 
 ## Source
diff --git a/docs/concepts/oracle-26ai.md b/docs/concepts/oracle-26ai.md
new file mode 100644
index 00000000..e0705b2f
--- /dev/null
+++ b/docs/concepts/oracle-26ai.md
@@ -0,0 +1,85 @@
+---
+title: Oracle Database 26ai
+description: "Every Oracle Database 26ai primitive in Locus — vector search, in-database chunking and embeddings, table ingestion, durable and versioned threads, long-term memory and Oracle agent memory — with the notebook that runs each one."
+---
+
+# Oracle Database 26ai
+
+Locus treats Oracle Database 26ai as the place where an agent's state lives:
+vectors for retrieval, threads for continuity, memories for what it has
+learned, in the same database as the data it acts on. Each primitive is a
+first-class Locus class on the `oracledb` driver, with no LangChain layer in
+between, and it works with the same `Agent` as everything else.
+
+```bash
+pip install "locus-sdk[oci,oracledb]"
+```
+
+## What ships
+
+| Need | Class | Notebook |
+|---|---|---|
+| Vector search for RAG on native `VECTOR` columns | `OracleVectorStore` | [38 — Oracle 26ai RAG](../notebooks/notebook_38_oracle_26ai_rag.md) |
+| Ingest rows from your own tables as documents | `OracleADBLoader` | [39 — ADB loader](../notebooks/notebook_39_oracle_adb_loader.md) |
+| Chunk text inside the database | `OracleInDBChunker` | [40 — In-database chunker](../notebooks/notebook_40_oracle_indb_chunker.md) |
+| Embed inside the database with an ONNX model | `OracleInDBEmbeddings` | [41 — In-database embeddings](../notebooks/notebook_41_oracle_indb_embeddings.md) |
+| Durable conversation threads | `oracle_checkpointer()` | [42 — Oracle 26ai checkpointer](../notebooks/notebook_42_oracle_26ai_checkpointer.md) |
+| Every version of a thread, resume from any point | `OracleCheckpointSaver` | [43 — Versioned saver](../notebooks/notebook_43_oracle_versioned_saver.md) |
+| Long-term memory store (`BaseStore`) | `OracleStore` | [37 — Oracle store](../notebooks/notebook_37_oracle_store.md) |
+| Oracle's agent-memory service, in-database extraction and recall | `OracleAgentMemoryManager` | [44 — Oracle agent memory](../notebooks/notebook_44_oracle_agent_memory.md) |
+
+Every class has a synchronous twin (`OracleSyncVectorStore`,
+`OracleSyncInDBEmbeddings`, `OracleSyncInDBChunker`, `OracleSyncADBLoader`,
+`OracleSyncStore`, `OracleSyncCheckpointSaver`) for code that is not async.
+
+## Connecting
+
+All of them take the same connection arguments: `dsn`, `user`, `password`,
+and for Autonomous Database `wallet_location` and `wallet_password`; or
+`host`, `port` and `service_name` for a plain listener. The notebooks and the
+integration tests read them from `ORACLE_DSN`, `ORACLE_USER`,
+`ORACLE_PASSWORD`, `ORACLE_WALLET` and `ORACLE_WALLET_PASSWORD`.
+
+## Retrieval
+
+The agent's own documents on a `VECTOR` column, searched with
+`VECTOR_DISTANCE`, exposed to the model as a tool.
+
+```python
+import os
+
+from locus.agent import Agent
+from locus.rag import OCIEmbeddings, OracleVectorStore, RAGRetriever
+
+db = dict(
+    dsn=os.environ["ORACLE_DSN"],
+    user=os.environ["ORACLE_USER"],
+    password=os.environ["ORACLE_PASSWORD"],
+    wallet_location=os.environ["ORACLE_WALLET"],
+    wallet_password=os.environ.get("ORACLE_WALLET_PASSWORD", ""),
+)
+retriever = RAGRetriever(
+    embedder=OCIEmbeddings(model_id="cohere.embed-v4.0"),
+    store=OracleVectorStore(table_name="runbooks", dimension=1536, **db),
+)
+agent = Agent(model="oci:openai.gpt-5.5", tools=[retriever.as_tool()])
+```
+
+When the text must not leave the database, chunk and embed inside it:
+`OracleInDBChunker` splits with `DBMS_VECTOR_CHAIN`, `OracleInDBEmbeddings`
+runs an ONNX model loaded into the database, and `OracleVectorStore` keeps
+the result, so no document content goes to a hosted embedding API.
+
+## Threads and memory
+
+`oracle_checkpointer()` gives every thread a durable home;
+`OracleCheckpointSaver` keeps the full history of each thread; `OracleStore`
+and `OracleAgentMemoryManager` hold what the agent learns across threads.
+[Memory: short-term and long-term](memory.md) explains the three layers
+and shows each one on Oracle Database 26ai.
+
+## Where to go next
+
+- [OCI Generative AI](providers/oci.md) for the model side of the same stack.
+- [Dedicated AI Clusters](../how-to/oci-dac.md) for your own model endpoints.
+- [RAG](rag.md), [Checkpointers](checkpointers.md) and [Long-term memory manager](memory-manager.md) for the general surfaces these classes plug into.
diff --git a/docs/concepts/prompts.md b/docs/concepts/prompts.md
index 616f7112..c551936f 100644
--- a/docs/concepts/prompts.md
+++ b/docs/concepts/prompts.md
@@ -103,7 +103,7 @@ charges fewer tokens on cache hits — pass `prompt_cache=True`:
 from locus.models import AnthropicModel
 
 agent = Agent(
-    model=AnthropicModel(model="claude-sonnet-4-6", prompt_cache=True),
+    model=AnthropicModel(model="claude-sonnet-5-5", prompt_cache=True),
     system_prompt=very_long_prompt,
 )
 ```
diff --git a/docs/concepts/providers/anthropic.md b/docs/concepts/providers/anthropic.md
index 6676652e..4e4c6e36 100644
--- a/docs/concepts/providers/anthropic.md
+++ b/docs/concepts/providers/anthropic.md
@@ -33,12 +33,12 @@ export ANTHROPIC_API_KEY=sk-ant-...
 ```python
 from locus.agent import Agent
 agent = Agent(
-    model="anthropic:claude-sonnet-4-6",
+    model="anthropic:claude-sonnet-5-5",
     system_prompt="You are a support agent.",
 )
 ```
 
-The string `"anthropic:claude-sonnet-4-6"` tells the SDK the
+The string `"anthropic:claude-sonnet-5-5"` tells the SDK the
 provider (`anthropic:`) and the exact model id. Any model id
 Anthropic accepts, the SDK accepts.
 
@@ -62,10 +62,10 @@ examples, not a maintained roster:
 
 | Model | When to pick it |
 |---|---|
-| `claude-opus-4-8` | Hardest problems — timeline reconstruction, complex multi-step plans |
-| `claude-sonnet-5` | Current-generation everyday workhorse |
-| `claude-sonnet-4-6` | Everyday workhorse — fast enough, smart enough, cheap enough for routine work |
-| `claude-haiku-4-5` | High-volume classification and routing — cheap calls, log summaries |
+| `claude-fable-5-1` | Hardest problems — long multi-step plans, timeline reconstruction |
+| `claude-opus-5-5` | Complex reasoning where cost matters less than quality |
+| `claude-sonnet-5-5` | Everyday workhorse — fast enough, smart enough, cheap enough for routine work |
+| `claude-haiku-5-5` | High-volume classification and routing — cheap calls, log summaries |
 
 ### Real SSE streaming
 
@@ -103,7 +103,7 @@ class Triage(BaseModel):
     needs_human: bool
 
 agent = Agent(
-    model="anthropic:claude-sonnet-4-6",
+    model="anthropic:claude-sonnet-5-5",
     output_schema=Triage,
 )
 result = agent.run_sync("Customer 4821 reports a duplicate charge on ord-4821.")
@@ -124,10 +124,9 @@ Opt in with `prompt_cache=True` on `AnthropicModel`. The SDK then sends
 the system prompt as a block list with `cache_control: ephemeral` and
 tags the last entry of the tool catalog the same way (Anthropic walks
 markers in order — the last tag anchors the cache point).
-These markers are applied on the non-streaming path only:
-`AnthropicModel.stream()`, which `agent.run(..., stream_tokens=True)`
-uses, sends the system prompt and tool catalog untagged, so nothing is
-cached while streaming tokens.
+The same markers go on the streaming path: `AnthropicModel.stream()`,
+which `agent.run(..., stream_tokens=True)` uses, builds its request through
+the same code as `complete()`, so streaming turns hit the cache too.
 
 ```python
 from locus.agent import Agent
@@ -135,7 +134,7 @@ from locus.models.native.anthropic import AnthropicModel
 
 agent = Agent(
     model=AnthropicModel(
-        model="claude-sonnet-4-6",
+        model="claude-sonnet-5-5",
         prompt_cache=True,
     ),
     tools=[...],
@@ -189,7 +188,7 @@ async for event in agent.run("...", stream_tokens=True):
 | `429 overloaded_error` | Anthropic capacity; the `ModelRetryHook` re-tries with backoff if installed |
 | Prompt caching not visible in usage stats | Cache window expired (5 min ephemeral) or prompt below the threshold |
 | `ThinkEvent`s never fire | The turn produced no assistant `message.content` (e.g. tool-call-only turn) |
-| `temperature` appears to have no effect | Anthropic deprecated `temperature` from `claude-opus-4-7` onward and across the Claude 5 family. The provider matches those ids by prefix (`claude-opus-4-7`, `claude-opus-4-8`, `claude-opus-4-9`, `claude-sonnet-5`, `claude-opus-5`, `claude-haiku-5`, `claude-fable-5`, `claude-mythos-5`) and omits the parameter rather than taking a 400 — it is dropped silently, with no warning, even though `AnthropicConfig.temperature` defaults to `0.7`. Separately, token streaming (`agent.run(..., stream_tokens=True)`) goes through `AnthropicModel.stream()`, which sends no `temperature` for any Claude model, so while streaming the setting has no effect on `claude-sonnet-4-6` or `claude-haiku-4-5` either. Steer with the system prompt instead. |
+| `temperature` appears to have no effect | Anthropic deprecated `temperature` from `claude-opus-4-7` onward and across the Claude 5 family. The provider matches those ids by prefix (`claude-opus-4-7`, `claude-opus-4-8`, `claude-opus-4-9`, `claude-sonnet-5`, `claude-opus-5`, `claude-haiku-5`, `claude-fable-5`, `claude-mythos-5`) and omits the parameter rather than taking a 400 — it is dropped silently, with no warning, even though `AnthropicConfig.temperature` defaults to `0.7`. Separately, token streaming (`agent.run(..., stream_tokens=True)`) goes through `AnthropicModel.stream()`, which sends no `temperature` for any Claude model, so while streaming the setting has no effect on `claude-sonnet-5-5` or `claude-haiku-4-5` either. Steer with the system prompt instead. |
 
 ## Source
 
diff --git a/docs/concepts/providers/resilience.md b/docs/concepts/providers/resilience.md
index 19db1b68..c0ff7ebb 100644
--- a/docs/concepts/providers/resilience.md
+++ b/docs/concepts/providers/resilience.md
@@ -55,7 +55,7 @@ pool = CredentialPool(
 
 def build(credential: Credential) -> AnthropicModel:
     return AnthropicModel(
-        model="claude-sonnet-4-6",
+        model="claude-sonnet-5-5",
         api_key=credential.api_key.get_secret_value(),
     )
 
diff --git a/docs/concepts/safety.md b/docs/concepts/safety.md
index 8f30177a..b266c006 100644
--- a/docs/concepts/safety.md
+++ b/docs/concepts/safety.md
@@ -123,7 +123,7 @@ from locus.hooks.builtin.steering import SteeringHook
 
 # The steering model is a model instance — it can be a smaller/cheaper
 # model than the agent's main one.
-judge = AnthropicModel(model="claude-sonnet-4-6")
+judge = AnthropicModel(model="claude-sonnet-5-5")
 
 agent = Agent(
     model="{{ locus_example_model }}",
diff --git a/docs/how-to/environment-variables.md b/docs/how-to/environment-variables.md
index b1f27c97..d8e07fb1 100644
--- a/docs/how-to/environment-variables.md
+++ b/docs/how-to/environment-variables.md
@@ -26,7 +26,7 @@ the notebooks run offline against a bundled model — setting only
 | Variable | Used by |
 |---|---|
 | `LOCUS_MODEL_PROVIDER` | `examples/config.py:get_model` — `oci` (auto-detected when `~/.oci/config` has a usable profile), `mock` (default otherwise), `openai`, `anthropic` |
-| `LOCUS_MODEL_ID` | Model id for the selected provider (e.g. `gpt-4o`, `claude-sonnet-4-6`) |
+| `LOCUS_MODEL_ID` | Model id for the selected provider (e.g. `gpt-4o`, `claude-sonnet-5-5`) |
 | `LOCUS_MODEL_ID_B` | Secondary "model B" slot — comparison notebooks |
 | `LOCUS_MODEL_ID_C` | Tertiary "model C" slot |
 | `LOCUS_RESEARCH_MODEL` | Overrides the model for the research/eval scripts |
diff --git a/docs/how-to/litellm-gateway.md b/docs/how-to/litellm-gateway.md
index 6dc7777f..eaf52fc0 100644
--- a/docs/how-to/litellm-gateway.md
+++ b/docs/how-to/litellm-gateway.md
@@ -71,7 +71,7 @@ if any of them is unset.
 
 The gateway listens on `http://localhost:4000` and exposes the model
 aliases declared in `config.yaml`. The sample ships six:
-`gpt-4o`, `gpt-4o-mini`, `gpt-5-mini`, `claude-sonnet-4-6`,
+`gpt-4o`, `gpt-4o-mini`, `gpt-5-mini`, `claude-sonnet-5-5`,
 `claude-haiku`, and `text-embedding-3-small`. Add more by extending
 `model_list`.
 
@@ -281,7 +281,7 @@ Highlights:
 - **`general_settings.master_key`** — the admin key that creates
   per-team virtual keys via `/key/generate`.
 - **`router_settings.fallbacks`** — fallback chains across model
-  aliases (e.g. `[{"gpt-5-mini": ["claude-sonnet-4-6"]}]`).
+  aliases (e.g. `[{"gpt-5-mini": ["claude-sonnet-5-5"]}]`).
 - **`litellm_settings.callbacks`** — observability hooks (Langfuse,
   OTel, Datadog, …).
 - **`litellm_settings.cache`** — Redis / S3 / Qdrant caching config.
diff --git a/docs/index.md b/docs/index.md
index 13d3a261..858a40c2 100644
--- a/docs/index.md
+++ b/docs/index.md
@@ -8,8 +8,6 @@ description: "Oracle Generative AI Agentic Harness. Locus is a Python agent harn
 <div class="locus-hero" markdown>
 <div class="locus-hero__copy" markdown>
 
-<p class="locus-product-name"><span class="locus-product-name__brand">Oracle Generative AI</span><span class="locus-product-name__name">Agentic Harness</span></p>
-
 # Multi-agent workflows built for <span class="accent">production.</span>
 
 One `Agent` class, eight ways to combine agents, Oracle Cloud underneath, and the runtime around the model: tools, memory, RAG, observability, guardrails, an admission gate.
@@ -77,7 +75,7 @@ print(agent.run_sync("What is the capital of France?").text)
 # → Paris
 ```
 
-Swap `oci:openai.gpt-5.5` for `openai:gpt-4o` or `anthropic:claude-sonnet-4-6`;
+Swap `oci:openai.gpt-5.5` for `openai:gpt-4o` or `anthropic:claude-sonnet-5-5`;
 the call stays the same. With an OCI profile in `~/.oci/config`, every notebook
 runs against OCI Generative AI with no further setup.
 
@@ -228,6 +226,7 @@ retriever = RAGRetriever(                            # Oracle Database 26ai
 agent = Agent(model="oci:openai.gpt-5.5", tools=[retriever.as_tool()])
 ```
 
+[Oracle Database 26ai, every primitive →](concepts/oracle-26ai.md) ·
 [Which OCI models work →](how-to/oci-models.md#model-support-on-oci) ·
 [Dedicated AI Clusters →](how-to/oci-dac.md) ·
 [Oracle Database 26ai notebooks →](notebooks/index.md#oracle-cloud-and-database-26ai)
@@ -242,11 +241,12 @@ agent = Agent(model="oci:openai.gpt-5.5", tools=[retriever.as_tool()])
     `@tool` on any function, typed arguments coerced from whatever the
     model sends, idempotent execution, MCP servers in or out.
 
-- :material-database:{ .lg .middle } **[Memory and RAG](concepts/checkpointers.md)**
+- :material-database:{ .lg .middle } **[Memory and RAG](concepts/memory.md)**
 
     ---
-    Durable threads on 10 backends, long-term memory, six vector stores,
-    rerankers, and `retriever.as_tool()`.
+    Four layers: working memory, the thread (10 checkpoint backends),
+    long-term memory across threads, and knowledge through RAG (six vector
+    stores, rerankers, `retriever.as_tool()`). All four on Oracle Database 26ai.
 
 - :material-shield-lock:{ .lg .middle } **[Grounding, guardrails, admission gate](concepts/control-layer.md)**
 
diff --git a/docs/stylesheets/locus.css b/docs/stylesheets/locus.css
index dda12d39..54304e59 100644
--- a/docs/stylesheets/locus.css
+++ b/docs/stylesheets/locus.css
@@ -388,47 +388,6 @@ html {
   border-radius: 50%;
   pointer-events: none;
 }
-/* Hero eyebrow — a pill badge: red dot · "Oracle Generative AI" (ink) ·
-   hairline · "Agentic Harness" (Oracle red). Compound
-   `.locus-hero p.locus-product-name` (specificity 0,3,1) beats
-   `.md-typeset .locus-hero p` (0,2,1), so the eyebrow size actually wins
-   inside the hero. Second selector covers usages outside the hero. */
-.md-typeset .locus-hero p.locus-product-name,
-.md-typeset p.locus-product-name {
-  display: inline-flex;
-  align-items: center;
-  gap: 0.6rem;
-  margin: 0 0 1.1rem;
-  padding: 0.34rem 0.9rem 0.34rem 0.75rem;
-  border: 1px solid color-mix(in srgb, var(--tl-red) 36%, transparent);
-  border-radius: 999px;
-  background: color-mix(in srgb, var(--tl-red) 8%, transparent);
-  font-family: var(--locus-sans);
-  font-size: 0.66rem !important;
-  line-height: 1.2;
-  font-weight: 600;
-  letter-spacing: 0.1em;
-  text-transform: uppercase;
-  color: var(--locus-ink);
-  opacity: 1;
-}
-.md-typeset p.locus-product-name::before {
-  content: "";
-  flex: none;
-  width: 0.45rem;
-  height: 0.45rem;
-  border-radius: 50%;
-  background: var(--tl-red);
-  box-shadow: 0 0 0 3px color-mix(in srgb, var(--tl-red) 22%, transparent);
-}
-.md-typeset p.locus-product-name .locus-product-name__brand {
-  opacity: 0.78;
-}
-.md-typeset p.locus-product-name .locus-product-name__name {
-  color: var(--tl-red);
-  padding-left: 0.6rem;
-  border-left: 1px solid color-mix(in srgb, var(--tl-red) 40%, transparent);
-}
 .md-typeset .locus-hero h1:first-of-type {
   font-family: var(--locus-sans);
   font-size: 2.9rem;
@@ -985,11 +944,6 @@ form.md-header__option {
   margin-top: 0;
   padding-top: 0.2rem;
 }
-[data-md-color-scheme="slate"] .md-nav--primary .md-nav__item--section > .md-nav__link,
-[data-md-color-scheme="slate"] .md-nav--primary .md-nav__item--section > .md-nav__link:not(.md-nav__link--active) {
-  color: var(--tl-rose-tint) !important;
-  border-top-color: var(--tl-rule);
-}
 
 .md-nav__link--active,
 .md-nav__link[aria-current="true"] {
@@ -1637,24 +1591,6 @@ form.md-header__option {
   }
   /* The one-line pill is ~400px; on phones stack brand over name inside a
      rounded tag instead of overflowing the column. */
-  .md-typeset .locus-hero p.locus-product-name,
-  .md-typeset p.locus-product-name {
-    display: inline-grid;
-    grid-template-columns: auto 1fr;
-    column-gap: 0.6rem;
-    row-gap: 0.15rem;
-    align-items: center;
-    border-radius: 0.75rem;
-    font-size: 0.6rem !important;
-    letter-spacing: 0.06em;
-  }
-  .md-typeset p.locus-product-name::before {
-    grid-row: 1 / span 2;
-  }
-  .md-typeset p.locus-product-name .locus-product-name__name {
-    padding-left: 0;
-    border-left: 0;
-  }
   /* At 0.78rem the one-liner spills past the copy button; at 0.7rem the
      whole command fits a 390px column with no scroll. */
   .md-typeset .locus-hero__copy pre {
diff --git a/examples/README.md b/examples/README.md
index 7bb44732..42f5ab25 100644
--- a/examples/README.md
+++ b/examples/README.md
@@ -52,7 +52,7 @@ need no provider at all.
 from locus.agent import Agent
 from locus.models import get_model
 
-model = get_model("openai:gpt-4o")  # or "anthropic:claude-sonnet-4-6"
+model = get_model("openai:gpt-4o")  # or "anthropic:claude-sonnet-5-5"
 
 agent = Agent(
     model=model,
diff --git a/examples/config.py b/examples/config.py
index 4b9d9f83..27e1482d 100644
--- a/examples/config.py
+++ b/examples/config.py
@@ -636,7 +636,7 @@ def _get_anthropic_model(**kwargs: Any) -> Any:
     """Get Anthropic model."""
     from locus.models.native.anthropic import AnthropicModel
 
-    model_id = kwargs.pop("model_id", os.environ.get("LOCUS_MODEL_ID", "claude-sonnet-4-6"))
+    model_id = kwargs.pop("model_id", os.environ.get("LOCUS_MODEL_ID", "claude-sonnet-5-5"))
     api_key = os.environ.get("ANTHROPIC_API_KEY")
 
     if not api_key:
diff --git a/examples/notebook_11_model_providers.py b/examples/notebook_11_model_providers.py
index 96838bab..3cb230bd 100644
--- a/examples/notebook_11_model_providers.py
+++ b/examples/notebook_11_model_providers.py
@@ -4,9 +4,9 @@
 """Notebook 11: Model providers — one cloud-ops agent, swap models with a string.
 
 The Gateway is the seam every AI request in the org flows through: one
-cloud-operations agent codebase, any backing model. Locus supports OpenAI
-and Anthropic as first-class providers, and the same agent code runs
-against either — only the model object changes, so a platform team can
+cloud-operations agent codebase, any backing model. Locus supports OCI
+Generative AI, OpenAI and Anthropic as first-class providers, and the same
+agent code runs against any of them — only the model object changes, so a platform team can
 re-platform a provider (for cost, latency, data residency, or an outage
 failover) without rewriting a single runbook. The provider abstraction is
 also a control point: routing, rate limits, and audit live at the Gateway,
@@ -14,8 +14,9 @@
 
 Provider matrix:
 
-  OpenAI    — GPT-4o, o1, o3, gpt-5.x against the direct API.
-  Anthropic — Claude models (opus / sonnet / haiku).
+  OCI       — oci:<vendor>.<model>: the OCI Generative AI catalog, auto-routed.
+  OpenAI    — GPT-4o, o3, gpt-5.x against the direct API.
+  Anthropic — Claude models (fable / opus / sonnet / haiku).
   Registry  — get_model("provider:model_name") returns the right client.
 
 Run it
@@ -67,7 +68,7 @@ async def example_providers():
 
     print("\nUsage:")
     print('  model = get_model("openai:gpt-4o")')
-    print('  model = get_model("anthropic:claude-sonnet-4-6")')
+    print('  model = get_model("anthropic:claude-sonnet-5-5")')
     print()
     print("The provider prefix before the colon selects the client; the rest")
     print("is the model id that provider expects. See docs/concepts/models.md.")
@@ -89,7 +90,7 @@ async def example_direct():
 
     print("\nAnthropic (requires ANTHROPIC_API_KEY):")
     print("  from locus.models.native.anthropic import AnthropicModel")
-    print('  model = AnthropicModel(model="claude-sonnet-4-6")')
+    print('  model = AnthropicModel(model="claude-sonnet-5-5")')
 
 
 async def example_live_call() -> None:
diff --git a/examples/notebook_67_litellm_gateway_cost.py b/examples/notebook_67_litellm_gateway_cost.py
index 10b4129b..f58d5b3c 100644
--- a/examples/notebook_67_litellm_gateway_cost.py
+++ b/examples/notebook_67_litellm_gateway_cost.py
@@ -191,7 +191,7 @@ def main() -> None:
     # ----- Step 1: issue two virtual keys, one per payments team -----------
     fraud_ops_key = issue_virtual_key(url, master_key, team="fraud-ops", models=["gpt-4o"])
     dispute_ops_key = issue_virtual_key(
-        url, master_key, team="dispute-ops", models=["gpt-4o", "claude-sonnet-4-6"]
+        url, master_key, team="dispute-ops", models=["gpt-4o", "claude-sonnet-5-5"]
     )
     print(" Virtual keys issued (each with a $5 hard budget):")
     print(f"   fraud-ops   (gpt-4o only):    {fraud_ops_key[:24]}...")
diff --git a/examples/projects/a2a-mesh/src/a2a_mesh/_model.py b/examples/projects/a2a-mesh/src/a2a_mesh/_model.py
index 0ea24cba..921baacb 100644
--- a/examples/projects/a2a-mesh/src/a2a_mesh/_model.py
+++ b/examples/projects/a2a-mesh/src/a2a_mesh/_model.py
@@ -70,7 +70,7 @@ def _openai_model() -> Any:
 def _anthropic_model() -> Any:
     from locus.models.native.anthropic import AnthropicModel
 
-    return AnthropicModel(model=os.environ.get("LOCUS_MODEL_ID", "claude-sonnet-4-6"))
+    return AnthropicModel(model=os.environ.get("LOCUS_MODEL_ID", "claude-sonnet-5-5"))
 
 
 def get_model() -> Any:
diff --git a/examples/projects/deep-research/README.md b/examples/projects/deep-research/README.md
index bbae6e41..71857861 100644
--- a/examples/projects/deep-research/README.md
+++ b/examples/projects/deep-research/README.md
@@ -51,7 +51,7 @@ export OPENSEARCH_USERNAME=<your-username>
 export OPENSEARCH_PASSWORD='...'
 export OPENAI_API_KEY=sk-...
 export ANTHROPIC_API_KEY=sk-ant-...
-export LOCUS_RESEARCH_MODEL=anthropic:claude-sonnet-4-6
+export LOCUS_RESEARCH_MODEL=anthropic:claude-sonnet-5-5
 
 python examples/projects/deep-research/demo_opensearch_multi_index.py
 ```
diff --git a/examples/projects/deep-research/demo_hello_world.py b/examples/projects/deep-research/demo_hello_world.py
index 04d77ed6..284c6488 100755
--- a/examples/projects/deep-research/demo_hello_world.py
+++ b/examples/projects/deep-research/demo_hello_world.py
@@ -9,7 +9,7 @@
 call both tools in one turn.
 
 Run:
-    LOCUS_MODEL=anthropic:claude-sonnet-4-6 \\
+    LOCUS_MODEL=anthropic:claude-sonnet-5-5 \\
     ANTHROPIC_API_KEY=sk-ant-... \\
     python examples/projects/deep-research/demo_hello_world.py
 """
@@ -37,7 +37,7 @@ def fetch_advisory(vendor: str) -> str:
 
 
 def main() -> int:
-    model_id = os.environ.get("LOCUS_MODEL", "anthropic:claude-sonnet-4-6")
+    model_id = os.environ.get("LOCUS_MODEL", "anthropic:claude-sonnet-5-5")
 
     print("== locus hello-world deepagent ==")
     print(f"   model      : {model_id}")
diff --git a/examples/projects/deep-research/demo_opensearch_multi_index.py b/examples/projects/deep-research/demo_opensearch_multi_index.py
index 97763bd9..ebc3d041 100755
--- a/examples/projects/deep-research/demo_opensearch_multi_index.py
+++ b/examples/projects/deep-research/demo_opensearch_multi_index.py
@@ -102,7 +102,7 @@ async def main() -> int:
     password = os.environ["OPENSEARCH_PASSWORD"]
     verify_certs = os.environ.get("OPENSEARCH_VERIFY_CERTS", "true").lower() == "true"
 
-    model_id = os.environ.get("LOCUS_RESEARCH_MODEL", "anthropic:claude-sonnet-4-6")
+    model_id = os.environ.get("LOCUS_RESEARCH_MODEL", "anthropic:claude-sonnet-5-5")
 
     run_id = uuid.uuid4().hex[:8].lower()
     intel_index = f"locus_intel_{run_id}"
diff --git a/examples/projects/deep-research/demo_smoke.py b/examples/projects/deep-research/demo_smoke.py
index cb41ca2a..ac30ce34 100755
--- a/examples/projects/deep-research/demo_smoke.py
+++ b/examples/projects/deep-research/demo_smoke.py
@@ -89,7 +89,7 @@ async def main() -> None:
     from locus.models import get_model
 
     chat_model = get_model(
-        "anthropic:claude-sonnet-4-6",  # reliable tool caller for smoke
+        "anthropic:claude-sonnet-5-5",  # reliable tool caller for smoke
     )
 
     agent = create_deepagent(
diff --git a/mkdocs.yml b/mkdocs.yml
index c6bd2b6e..02393770 100644
--- a/mkdocs.yml
+++ b/mkdocs.yml
@@ -163,6 +163,7 @@ nav:
     - A2A protocol: concepts/multi-agent/a2a.md
     - DeepAgent: concepts/deepagent.md
     - Production-readiness: concepts/multi-agent/production.md
+  - Oracle Database 26ai: concepts/oracle-26ai.md
   - Agents:
     - State: concepts/state.md
     - Prompts: concepts/prompts.md
@@ -179,6 +180,7 @@ nav:
     - MCP: concepts/mcp.md
     - Multi-modal providers: concepts/multi-modal-providers.md
   - Memory and RAG:
+    - Memory — short-term and long-term: concepts/memory.md
     - Checkpointers & Store: concepts/checkpointers.md
     - Long-term memory manager: concepts/memory-manager.md
     - RAG: concepts/rag.md
diff --git a/src/locus/__init__.py b/src/locus/__init__.py
index d4387679..031f5854 100644
--- a/src/locus/__init__.py
+++ b/src/locus/__init__.py
@@ -22,7 +22,7 @@ def search(query: str) -> str:
         return "results..."
 
     agent = Agent(
-        model="openai:gpt-4o",  # or anthropic:claude-sonnet-4-6
+        model="openai:gpt-4o",  # or anthropic:claude-sonnet-5-5
         tools=[search],
         system_prompt="You are a helpful assistant.",
     )
diff --git a/src/locus/agent/agent.py b/src/locus/agent/agent.py
index 465959ad..2a8b66bd 100644
--- a/src/locus/agent/agent.py
+++ b/src/locus/agent/agent.py
@@ -118,7 +118,7 @@ class Agent(AgentRuntimeMixin, BaseModel):
 
     Usage:
         agent = Agent(
-            model="openai:gpt-4o",  # or anthropic:claude-sonnet-4-6
+            model="openai:gpt-4o",  # or anthropic:claude-sonnet-5-5
             tools=[search, calculate],
             system_prompt="You are a helpful assistant.",
         )
diff --git a/src/locus/agent/config.py b/src/locus/agent/config.py
index 8c8779c2..120d7b68 100644
--- a/src/locus/agent/config.py
+++ b/src/locus/agent/config.py
@@ -291,7 +291,7 @@ class AgentConfig(BaseModel):
     # Model specification
     model: str | Any = Field(
         ...,
-        description="Model string ('openai:gpt-4o' or 'anthropic:claude-sonnet-4-6') or ModelProtocol instance",
+        description="Model string ('openai:gpt-4o' or 'anthropic:claude-sonnet-5-5') or ModelProtocol instance",
     )
 
     # Provider configuration that has to travel with the string form.
diff --git a/src/locus/agent/specs.py b/src/locus/agent/specs.py
index d174c425..b4ed2c58 100644
--- a/src/locus/agent/specs.py
+++ b/src/locus/agent/specs.py
@@ -14,7 +14,7 @@
     ---
     name: reviewer
     description: Reviews a diff for bugs. Use after a change, before commit.
-    model: anthropic:claude-sonnet-4-6
+    model: anthropic:claude-sonnet-5-5
     tools: read, grep, glob          # or a YAML list
     disallowed_tools: [bash]
     mode: subagent                   # primary | subagent | all
diff --git a/src/locus/control/verification.py b/src/locus/control/verification.py
index e2784288..8647d32e 100644
--- a/src/locus/control/verification.py
+++ b/src/locus/control/verification.py
@@ -265,7 +265,7 @@ class AdversarialSkeptic:
             finding,
             skeptics=[
                 EvidenceQualitySkeptic(),
-                AdversarialSkeptic("anthropic:claude-sonnet-4-6"),
+                AdversarialSkeptic("anthropic:claude-sonnet-5-5"),
             ],
         )
     """
diff --git a/src/locus/memory/manager.py b/src/locus/memory/manager.py
index 7ceb1596..681598b4 100644
--- a/src/locus/memory/manager.py
+++ b/src/locus/memory/manager.py
@@ -88,7 +88,7 @@ async def my_extractor(messages):
 
 
     agent = Agent(
-        model="anthropic:claude-sonnet-4-6",
+        model="anthropic:claude-sonnet-5-5",
         memory_manager=LLMMemoryManager(store=store, extract_fn=my_extractor),
     )
 """
@@ -528,7 +528,7 @@ class LLMMemoryManager(BaseMemoryManager):
             namespace_prefix=("users", user_id),
         )
 
-        agent = Agent(model="anthropic:claude-sonnet-4-6", memory_manager=manager)
+        agent = Agent(model="anthropic:claude-sonnet-5-5", memory_manager=manager)
     """
 
     def __init__(
diff --git a/src/locus/memory/managers/mem0.py b/src/locus/memory/managers/mem0.py
index b500168a..14ae3136 100644
--- a/src/locus/memory/managers/mem0.py
+++ b/src/locus/memory/managers/mem0.py
@@ -20,7 +20,7 @@
     from locus.memory.managers import Mem0MemoryManager
 
     manager = Mem0MemoryManager(user_id="alice")
-    agent = Agent(model="anthropic:claude-sonnet-4-6", memory_manager=manager)
+    agent = Agent(model="anthropic:claude-sonnet-5-5", memory_manager=manager)
 
 By default mem0 builds its own OpenAI-backed LLM + embedder + local
 vector store; pass ``config=`` to point it at other backends (a fully
diff --git a/src/locus/models/__init__.py b/src/locus/models/__init__.py
index 58513a25..2ded1de5 100644
--- a/src/locus/models/__init__.py
+++ b/src/locus/models/__init__.py
@@ -21,7 +21,7 @@
     model = OpenAIModel(model="gpt-4o")
 
     from locus.models import AnthropicModel
-    model = AnthropicModel(model="claude-sonnet-4-6")
+    model = AnthropicModel(model="claude-sonnet-5-5")
 
     # OCI GenAI — V1 transport
     from locus.models import OCIChatCompletionsModel
@@ -29,7 +29,7 @@
 
     # String factory — "provider:model"
     from locus.models import get_model
-    model = get_model("anthropic:claude-sonnet-4-6")
+    model = get_model("anthropic:claude-sonnet-5-5")
     model = get_model("oci:openai.gpt-5", profile="DEFAULT")
 
 Provider failover — ``FallbackChain`` (see :mod:`locus.models.fallback`)::
@@ -37,7 +37,7 @@
     from locus.models import FallbackChain
 
     model = FallbackChain(
-        [get_model("anthropic:claude-sonnet-4-6"), get_model("openai:gpt-4o")]
+        [get_model("anthropic:claude-sonnet-5-5"), get_model("openai:gpt-4o")]
     )
 """
 
diff --git a/src/locus/models/fallback.py b/src/locus/models/fallback.py
index 96018a30..18dc15b1 100644
--- a/src/locus/models/fallback.py
+++ b/src/locus/models/fallback.py
@@ -14,7 +14,7 @@
 
     model = FallbackChain(
         [
-            get_model("anthropic:claude-sonnet-4-6"),
+            get_model("anthropic:claude-sonnet-5-5"),
             get_model("oci:meta.llama-3.3-70b-instruct"),
             get_model("openai:gpt-4o"),
         ],
diff --git a/src/locus/models/native/anthropic.py b/src/locus/models/native/anthropic.py
index 7655363f..24c66e4b 100644
--- a/src/locus/models/native/anthropic.py
+++ b/src/locus/models/native/anthropic.py
@@ -238,7 +238,7 @@ def _usage_dict(usage: Any) -> dict[str, int]:
 class AnthropicConfig(ModelConfig):
     """Configuration for Anthropic models."""
 
-    model: str = "claude-sonnet-4-6"
+    model: str = "claude-sonnet-5-5"
     max_tokens: int = 4096
     temperature: float = 0.7
     top_p: float = 0.9
@@ -302,10 +302,11 @@ class AnthropicConfig(ModelConfig):
 class AnthropicModel(BaseModel):
     """Anthropic model provider.
 
-    Supports Claude 4.6, 4.5, 3.5 models with streaming and tool calling.
+    Supports the Claude 5 family (Fable 5.1, Opus / Sonnet / Haiku 5.5) and
+    earlier Claude models, with streaming and tool calling.
 
     Example:
-        >>> model = AnthropicModel(model="claude-sonnet-4-6")
+        >>> model = AnthropicModel(model="claude-sonnet-5-5")
         >>> response = await model.complete([Message.user("Hello!")])
     """
 
@@ -325,7 +326,7 @@ def supports_structured_output(self) -> bool:
 
     def __init__(
         self,
-        model: str = "claude-sonnet-4-6",
+        model: str = "claude-sonnet-5-5",
         api_key: str | None = None,
         base_url: str | None = None,
         max_tokens: int = 4096,
diff --git a/src/locus/models/pooled.py b/src/locus/models/pooled.py
index 373e599d..3ab946df 100644
--- a/src/locus/models/pooled.py
+++ b/src/locus/models/pooled.py
@@ -26,7 +26,7 @@
 
     def _build(cred: Credential) -> AnthropicModel:
         return AnthropicModel(
-            model="claude-sonnet-4-6",
+            model="claude-sonnet-5-5",
             api_key=cred.api_key,
             ...
         )
diff --git a/src/locus/models/profiles.py b/src/locus/models/profiles.py
index 7caabc98..57cf8652 100644
--- a/src/locus/models/profiles.py
+++ b/src/locus/models/profiles.py
@@ -231,7 +231,7 @@ class ModelProfile(BaseModel):
     "default": {"prompt_variant": "default"},
 }
 
-#: A Claude id's tier and version: ``claude-sonnet-4-6`` → (4, 6),
+#: A Claude id's tier and version: ``claude-sonnet-5-5`` → (5, 5),
 #: ``claude-opus-5`` → (5, None), ``claude-sonnet-4-5-20250929`` → (4, 5).
 _CLAUDE_VERSION: Final[re.Pattern[str]] = re.compile(
     r"claude-(?:opus|sonnet|haiku|fable|mythos)-(\d+)(?:-(\d{1,2}))?(?:$|[-@.])"
diff --git a/src/locus/models/registry.py b/src/locus/models/registry.py
index c3784922..1aaf7eae 100644
--- a/src/locus/models/registry.py
+++ b/src/locus/models/registry.py
@@ -37,7 +37,7 @@ def get_model(model_string: str, **kwargs: Any) -> ModelProtocol:
 
     Examples:
         - "openai:gpt-4o"
-        - "anthropic:claude-sonnet-4-6"
+        - "anthropic:claude-sonnet-5-5"
 
     Args:
         model_string: Model identifier in "provider:model" format
diff --git a/src/locus/rogue/challenge.py b/src/locus/rogue/challenge.py
index 4b7b29b4..05da8842 100644
--- a/src/locus/rogue/challenge.py
+++ b/src/locus/rogue/challenge.py
@@ -265,7 +265,7 @@ def build_agent(mode: str) -> Agent:
             temperature=0,
         )
     elif mode == MODE_FRONTIER:
-        model = get_model("anthropic:claude-sonnet-4-6", api_key=os.environ["ANTHROPIC_API_KEY"])
+        model = get_model("anthropic:claude-sonnet-5-5", api_key=os.environ["ANTHROPIC_API_KEY"])
     else:
         model = CompromisedModel()
     return Agent(