feat: agent memory docs#30118
Conversation
- Add product directory entry (agent-memory.yaml) and SVG icon - Add product overview page with features, related products, and resource links - Add get-started guide: create namespace, configure binding, ingest/recall/summary - Add concepts: how Agent Memory works (extraction, classification, dedup, storage, recall pipeline, memory types, idempotency) and profiles/namespaces isolation model - Add configuration page for memory_namespaces Worker binding (single and multiple bindings, TypeScript types) - Add API reference: ingest() with Message type, timestamps, and idempotency; recall() with thinking levels, response length, and reference date; getSummary() with session scoping and summary structure - Add namespace management REST API reference (create, list, get, delete) - Add platform pages: validation limits and naming constraints, beta pricing placeholder - Add example: chat agent with memory using Workers AI (ingest, recall, getSummary in a conversation loop)
|
Hi @pmesgari — because this PR is from a fork, preview deploys don't run on this repo's CI. I've pushed the same commit to a branch on |
|
Moved to #30121 |
![devin-ai-integration[bot]](https://avatars.githubusercontent.com/in/811515?s=60&v=4){kind=small}
| -d '{"name": "my-app"}' | ||
| ``` | ||
|
|
||
| The response includes a `namespace_id` that you will use to configure your Worker binding. |
There was a problem hiding this comment.
🔴 Guide incorrectly tells users to use namespace_id instead of namespace name for binding configuration
The get-started guide tells users "The response includes a namespace_id that you will use to configure your Worker binding" (line 35) and "Replace the namespace value with the namespace_id from step 1" (line 62). However, the binding configuration placeholder is <NAMESPACE_NAME> (line 74), and src/content/docs/agent-memory/configuration/bindings.mdx:17 explicitly states "the namespace field is the namespace name from the Namespace Management API." This mismatch will cause users to put a ULID-style ID (like 01JSGC...) into the namespace field instead of the human-readable name (like my-app), resulting in a misconfigured binding.
| The response includes a `namespace_id` that you will use to configure your Worker binding. | |
| The response includes the namespace details. You will use the namespace `name` to configure your Worker binding. | |
Was this helpful? React with 👍 or 👎 to provide feedback.
|
|
||
| ## 3. Add the memory binding | ||
|
|
||
| Add the `agent_memory` binding to your Wrangler configuration. Replace the `namespace` value with the `namespace_id` from step 1. |
There was a problem hiding this comment.
🔴 Step 3 instruction incorrectly says to use namespace_id but placeholder and binding docs expect namespace name
Line 62 instructs users to "Replace the namespace value with the namespace_id from step 1." This contradicts the <NAMESPACE_NAME> placeholder on line 74 and the authoritative binding documentation at src/content/docs/agent-memory/configuration/bindings.mdx:17 which says the field takes the namespace name. Users following this instruction would use the wrong value.
| Add the `agent_memory` binding to your Wrangler configuration. Replace the `namespace` value with the `namespace_id` from step 1. | |
| Add the `agent_memory` binding to your Wrangler configuration. Replace the `namespace` value with the namespace name (`my-app`) from step 1. | |
Was this helpful? React with 👍 or 👎 to provide feedback.
|
|
||
| A namespace is a top-level container that scopes a set of memory profiles. You typically create one namespace per application, or one per environment (development, staging, production) within an application. | ||
|
|
||
| Namespaces are created through the [Namespace Management API](/agent-memory/api/namespaces/) and referenced by ID in your [Wrangler binding configuration](/agent-memory/configuration/bindings/). |
There was a problem hiding this comment.
🔴 Profiles-and-namespaces page says namespaces are "referenced by ID" but bindings use the name
Line 17 states namespaces are "referenced by ID in your Wrangler binding configuration," but the Wrangler binding namespace field takes the namespace name (e.g., my-app), not the ID (e.g., 01JSGCEXAMPLE000000000000). This contradicts src/content/docs/agent-memory/configuration/bindings.mdx:17 which says "the namespace field is the namespace name." All config examples across the docs consistently use <NAMESPACE_NAME> as the placeholder.
| Namespaces are created through the [Namespace Management API](/agent-memory/api/namespaces/) and referenced by ID in your [Wrangler binding configuration](/agent-memory/configuration/bindings/). | |
| Namespaces are created through the [Namespace Management API](/agent-memory/api/namespaces/) and referenced by name in your [Wrangler binding configuration](/agent-memory/configuration/bindings/). | |
Was this helpful? React with 👍 or 👎 to provide feedback.
| label: Guide | ||
| --- | ||
|
|
||
| import { TypeScriptExample, WranglerConfig, Steps } from "~/components"; |
There was a problem hiding this comment.
🟡 Unused Steps component imported in get-started guide
Line 12 imports Steps from ~/components, but the component is never used anywhere in the file. The guide uses manual ## 1. / ## 2. headings instead. This is dead code that should be removed for cleanliness.
| import { TypeScriptExample, WranglerConfig, Steps } from "~/components"; | |
| import { TypeScriptExample, WranglerConfig } from "~/components"; | |
Was this helpful? React with 👍 or 👎 to provide feedback.
3 participants
Summary
Adding docs for Agent Memory (Beta) product.
Screenshots (optional)
Documentation checklist