Add agent memories

README.md

@@ -54,7 +54,7 @@ Next, configure your LLM provider. ControlFlow's default provider is OpenAI, whi
 export OPENAI_API_KEY=your-api-key
 ```
 
-To use a different LLM provider, [see the LLM configuration docs](https://controlflow.ai/guides/llms).
+To use a different LLM provider, [see the LLM configuration docs](https://controlflow.ai/guides/configure-llms).
 
 
 ## Workflow Example

docs/concepts/agents.mdx

@@ -67,7 +67,7 @@ Tools are Python functions that the agent can call to perform specific actions o
 
 Each agent has a model, which is the LLM that powers the agent responses. This allows you to choose the most suitable model for your needs, based on factors such as performance, latency, and cost.
 
-ControlFlow supports any LangChain LLM that supports chat and function calling. For more details on how to configure models, see the [LLMs guide](/guides/llms).
+ControlFlow supports any LangChain LLM that supports chat and function calling. For more details on how to configure models, see the [LLMs guide](/guides/configure-llms).
 
 ### Interactivity
 

docs/guides/llms.mdx → docs/guides/configure-llms.mdx

@@ -1,5 +1,5 @@
 ---
-title: Configuring LLM models
+title: LLM Models
 description: ControlFlow supports a variety of LLMs and model providers.
 icon: sliders
 ---

docs/guides/default-agent.mdx

@@ -1,5 +1,6 @@
 ---
-title: Configuring the Default Agent
+title: Configuring a Default Agent
+sidebarTitle: Default Agent
 description: Set global and flow-specific defaults.
 icon: robot
 ---
@@ -8,7 +9,7 @@ ControlFlow uses a default agent when no specific agents are assigned to a task.
 
 ## Changing the Global Default Agent
 
-The global default agent (whose name, of course, is Marvin) uses whatever [default model](/guides/llms#changing-the-default-model) you've configured. It has a basic set of general-purpose instructions and is not equipped with any tools.
+The global default agent (whose name, of course, is Marvin) uses whatever [default model](/guides/configure-llms#changing-the-default-model) you've configured. It has a basic set of general-purpose instructions and is not equipped with any tools.
 
 To change the global default agent, assign a new agent to `controlflow.defaults.agent`:
 

docs/guides/default-memory.mdx

@@ -0,0 +1,60 @@
+---
+title: Configure a Default Memory Provider
+sidebarTitle: Default Memory Provider
+description: Set up a default persistent memory provider for your agents
+icon: brain
+---
+
+ControlFlow's [memory](/patterns/memory) feature allows agents to store and retrieve information across multiple workflows. Memory modules are backed by a vector database, configured using a `MemoryProvider`. 
+
+Setting up a default provider simplifies the process of creating memory objects throughout your application. Once configured, you can create memory objects without specifying a provider each time.
+
+<Tip>
+While ControlFlow does not include any vector database dependencies by default, the default provider is set to `"chroma-db"`. This means that if you install the `chromadb` package, your memory modules will work without any additional configuration.
+</Tip>
+
+## Install dependencies
+
+To use a provider, you must first install its dependencies. Please refer to the [Memory doc](/patterns/memory) to see all supported providers and their required dependencies. 
+
+For example, to use the default [Chroma](https://trychroma.com/) provider, you need to install `chromadb`:
+
+```bash
+pip install chromadb
+```
+
+## Configure a default provider
+
+There are two ways to set up a default provider: using a string setting for common defaults, or instantiating a custom provider. Here, we'll use a persistent Chroma database as our example.
+
+### String configurations
+
+For simple provider setups, you can modify ControlFlow's default settings using a string value. The default value is `"chroma-db"`, which will create a persistent Chroma database. To change it:
+
+<CodeGroup>
+```bash Environment variable
+export CONTROLFLOW_MEMORY_PROVIDER="chroma-ephemeral"
+```
+```python Runtime
+import controlflow as cf
+
+cf.settings.memory_provider = "chroma-ephemeral"
+```
+</CodeGroup>
+
+For a list of available string configurations, see the [Memory pattern guide](/patterns/memory).
+
+### Custom provider configuration
+
+For more advanced setups, instantiate a provider with custom settings and assign it to the ControlFlow default. Note this must be done at runtime.
+
+```python
+import controlflow as cf
+from controlflow.memory.providers.chroma import ChromaMemory
+import chromadb
+
+# Set the default provider
+cf.defaults.memory_provider = ChromaMemory(
+    client=chromadb.PersistentClient(path="/custom/path"),
+)
+```

docs/installation.mdx

@@ -35,7 +35,7 @@ export CONTROLFLOW_LLM_MODEL="anthropic/claude-3-5-sonnet-20240620"
 
 ### Other providers
 
-ControlFlow supports many other LLM providers as well, though you'll need to install their respective packages and configure the default LLM appropriately. See the [LLM documentation](/guides/llms) for more information.
+ControlFlow supports many other LLM providers as well, though you'll need to install their respective packages and configure the default LLM appropriately. See the [LLM documentation](/guides/configure-llms) for more information.
 
 
 ## Next steps

docs/mint.json

@@ -57,6 +57,7 @@
                 "patterns/tools",
                 "patterns/interactivity",
                 "patterns/dependencies",
+                "patterns/memory",
                 "patterns/instructions",
                 "patterns/planning",
                 "patterns/history"
@@ -66,8 +67,9 @@
             "group": "Configuration",
             "pages": [
                 "guides/settings",
-                "guides/llms",
-                "guides/default-agent"
+                "guides/configure-llms",
+                "guides/default-agent",
+                "guides/default-memory"
             ]
         },
         {