Add Redis Cloud Agent Memory REST quickstart

[n-valchev]

Summary

• Add a runnable Redis Cloud Agent Memory REST quickstart covering session events, session retrieval, long-term memory creation, and memory search.
• Connect the service creation and service details pages to the new workflow.
• Align the Agent Memory developer landing page, developer guide, and API examples with the quickstart.

Testing

• Full Hugo site build with garbage collection and debug logging
• Shell syntax validation for the runnable examples
• JSON request-body validation with jq
• Independent technical accuracy review against the Smithy API definitions and integration tests

---

Note

Low Risk
Hugo/markdown documentation only; no application code, APIs, or infrastructure changes.

Overview
Adds a Redis Cloud Agent Memory REST quickstart by turning use-agent-memory.md into an end-to-end tutorial: prerequisites, service creation, shell env vars (AGENT_MEMORY_URL, STORE_ID, API_KEY), and runnable curl flows for session events, session retrieval, direct long-term memory creation, and semantic search—with notes on async extraction and API key handling.

Recenters the doc set on that path. Landing cards on the Agent Memory index and developer guide now promote the REST quickstart; create-service, view-service, and the rc-agent-memory-get-started embed point readers to “continue with” or follow the quickstart instead of scattered API-only links.

Trims duplicate getting-started material from api-examples.md and developer-guide.md (inline auth curl, duplicated JSON quick starts) so those pages defer to the quickstart for setup and keep JSON-focused examples plus API reference links. Minor example cleanup includes aligning the long-term-memory path (no trailing slash) and standardizing env var naming.

^{Reviewed by Cursor Bugbot for commit d350227. Bugbot is set up for automated code reviews on this repo. Configure here.}

[github-actions]

Staging links:
https://redis.io/docs/staging/agent-memory-rest-quickstart/develop/ai/context-engine/agent-memory/
https://redis.io/docs/staging/agent-memory-rest-quickstart/develop/ai/context-engine/agent-memory/api-examples/
https://redis.io/docs/staging/agent-memory-rest-quickstart/develop/ai/context-engine/agent-memory/developer-guide/
https://redis.io/docs/staging/agent-memory-rest-quickstart/operate/rc/context-engine/agent-memory/
https://redis.io/docs/staging/agent-memory-rest-quickstart/operate/rc/context-engine/agent-memory/create-service/
https://redis.io/docs/staging/agent-memory-rest-quickstart/operate/rc/context-engine/agent-memory/use-agent-memory/
https://redis.io/docs/staging/agent-memory-rest-quickstart/operate/rc/context-engine/agent-memory/view-service/

[mich-elle-luna]

Thank you @n-valchev , just a few small items from claude tech accuracy review:

Issue 1 — "Quick create" description undersells the restriction (minor, worth clarifying)
The quickstart says:

Select Quick create to use the default settings, or select Create custom and choose an eligible database.

But create-service.md (the source of truth for that UI) says:

Select Quick create to create an Agent Memory service with default settings using your Free 30MB database. If you haven't created a Free database yet, Redis Cloud will create one for you.

A user who already has a Pro database and selects Quick create expecting it to use their Pro database will be surprised when it uses (or creates) a Free 30MB database instead. The quickstart should note that Quick create targets the Free database and that users with an existing database should use Create custom.

Issue 2 — Five-minute extraction cadence is a specific claim to flag for review (informational)
By default, extraction runs on a five-minute cadence, so extracted memories might not appear immediately.

This can't be verified from the OpenAPI spec. It's a correct-sounding implementation detail, but if the cadence changes in a future release it'll silently go stale. Worth flagging to the team so someone can confirm it's accurate and decide how precisely to state it ("may take several minutes" would age better than a hard number if the cadence is tunable or release-dependent).

Issue 3 — --fail-with-body requires curl 7.76.0+ (minor)
All four curl commands use --fail-with-body, which was added in curl 7.76.0 (April 2021). This is fine for any modern macOS or Linux system, but might silently fail with an "unknown option" error on older or minimal environments (Alpine-based containers, RHEL 7/8 with default curl). The older --fail flag is universally available — the only difference is that --fail-with-body still prints the response body on HTTP errors, which is helpful for debugging.

[mich-elle-luna]

thank you, just one more change to remove the duplicated note, otherwise this is looking good

[mich-elle-luna]

Suggested change

Redis Agent Memory on Redis Cloud is available as a public preview. Features and behavior can change before general availability.

[mich-elle-luna]

This note duplicates the banner we have on every page during the public preview.

[seth-riedel]

LGTM