developer-tools

Supermemory Review: AI Agent Memory Layer Tested (2026)

Name: Supermemory
Availability: InStock
Author: AI Demos

Hosted agent memory with strong capture and scoping, but mixed retrieval and weak forget behavior.

Visit Supermemory

API + SDKScoped by user/projectMemory GraphForget failure

TL;DR — our verdictUpdated July 2026 · 20 test artifacts

Useful for visible agent memory, but not for trusted forgetting

Where it wins

you need a hosted memory layer with API/SDK access and a visible UI for inspecting stored context
you need user, account, or project scoping to keep client and project histories separate
you want to test real multi-session memory workflows across preference, support, and project continuity

Main limitation

you need guaranteed forget/delete retirement to be correct on the first try

Pricing (verified plans)

Free $0Pro $19/moMax $100/moScale $399/mo

Strongest test artifacts

Preference Memory Persistence →Scope Isolation by Customer and Project →Support History Retrieval →

Our take

Supermemory is a solid hosted memory layer for AI agents when you need durable context capture, scoped user/project containers, and visible debugging. It remembered style, client, and project context well, and it kept separate containers isolated, but retrieval was inconsistent, the chat-level "0 memory used" tag conflicted with the side panel, and a forget request did not reliably retire the old preference. I would use it where observability matters, but not where delete/forget correctness is mission-critical.

Screen-recorded walkthrough used during testing.

In-Depth Review

Our detailed analysis of Supermemory — features, performance, and real-world testing.

AI Demos Team

Expert Reviewer

Verified Review

Feature-by-Feature Breakdown

Preference Memory Persistence

Mostly strong

▾

Test Summary

Feature tested: Preference Memory Persistence

Result: Partial — Mostly strong

Feature tested: Preference Memory Persistence

Result: Partial

Verdict: Mostly strong

Expected behavior: Stores a user's durable writing preferences and reuses them later. In this test it captured founder_001's short, direct, copy-paste-ready style, non-polished tone, proof-before-claim habit, and safer-next-step preference, then carried that context into an internal update.

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The assistant acknowledged the preference and stored a style-memory note that matched the short, direct working style. — Supermemory_Input1_Session1_Working_Style_Memory.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The assistant acknowledged the preference and stored a style-memory note that matched the short, direct working style. — Supermemory_Input1_Session1_Working_Style_Memory.png

What changed: Text prompt transformed into Image

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The internal update reflected the memory-testing context and stayed practical, though it was a little more polished and structured than the requested style. — Supermemory_Input1_Session2_Internal_Update_Output.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The internal update reflected the memory-testing context and stayed practical, though it was a little more polished and structured than the requested style. — Supermemory_Input1_Session2_Internal_Update_Output.png

What changed: Text prompt transformed into Image

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The assistant wrote a professional partner email without incorrectly forcing the internal-update tone into an unrelated communication. — Supermemory_Input1_Session3_Formal_Email_Output.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The assistant wrote a professional partner email without incorrectly forcing the internal-update tone into an unrelated communication. — Supermemory_Input1_Session3_Formal_Email_Output.png

What changed: Text prompt transformed into Image

Why it matters / Conclusion: Capture and reuse were strong, but the second-session output was slightly more polished than the user's stated preference.

Stores a user's durable writing preferences and reuses them later. In this test it captured founder_001's short, direct, copy-paste-ready style, non-polished tone, proof-before-claim habit, and safer-next-step preference, then carried that context into an internal update.

INPUT

Session 1 under user_id founder_001: I run a small AI product/research team. When you help me, remember how I work: keep outputs short, direct, and copy-paste ready; do not make writing sound too polished or motivational; always mention what proof or artifact is needed before making a strong claim; if a task is risky or unclear, tell me the safest next step instead of guessing.

↓→

image

Output artifact for "Preference Memory Persistence" test: The assistant acknowledged the preference and stored a style-memory note that matched the short, direct working style., Supermemory_Input1_Session1_Working_Style_Memory.png

The assistant acknowledged the preference and stored a style-memory note that matched the short, direct working style.

INPUT

Session 2 under user_id founder_001: Today I am testing tools for an AI memory use case. I want to show users that memory is not just 'remember my favorite color.' It should help an assistant continue real work across days, remember my working style, and avoid repeating the same explanation again. Create a short internal update for my team about what I worked on today and what we should test next.

↓→

image

Output artifact for "Preference Memory Persistence" test: The internal update reflected the memory-testing context and stayed practical, though it was a little more polished and structured than the requested style., Supermemory_Input1_Session2_Internal_Update_Output.png

The internal update reflected the memory-testing context and stayed practical, though it was a little more polished and structured than the requested style.

INPUT

Session 3 under user_id founder_001: Now write a formal email to a potential enterprise partner asking if they are open to a product demo next week. Keep it professional.

↓→

image

Output artifact for "Preference Memory Persistence" test: The assistant wrote a professional partner email without incorrectly forcing the internal-update tone into an unrelated communication., Supermemory_Input1_Session3_Formal_Email_Output.png

The assistant wrote a professional partner email without incorrectly forcing the internal-update tone into an unrelated communication.

Bottom Line

Capture and reuse were strong, but the second-session output was slightly more polished than the user's stated preference.

Scope Isolation by Customer and Project

Strong

▾

Test Summary

Feature tested: Scope Isolation by Customer and Project

Result: Passed — Strong

Keeps memories separated by customer/account and project container so one context does not leak into another. The benchmark kept ACME distinct from BetaCorp and separated the AI Demos memory use case from an unrelated sales-email-agent project.

INPUT

Session 1 under account_id client_acme_001: ACME is a client using our AI support assistant. They prefer clear next steps and do not like repeated troubleshooting. Their team already tried password reset, clearing browser cache, and switching browsers. The issue is still happening only for users with SSO enabled.

↓→

image

Output artifact for "Scope Isolation by Customer and Project" test: ACME support context was stored under client_acme_001 with its own troubleshooting history., Supermemory_Input2_Session1_ACME_Initial_SSO_Context.png

ACME support context was stored under client_acme_001 with its own troubleshooting history.

INPUT

Session 5 under account_id client_beta_002: BetaCorp is a new client. They say: 'Our users cannot log in for the first time.' Draft the first support reply for BetaCorp.

↓→

image

Output artifact for "Scope Isolation by Customer and Project" test: BetaCorp started as a fresh client container and did not inherit ACME's SSO history., Supermemory_Input2_Session5_BetaCorp_Scope_Isolation.png

BetaCorp started as a fresh client container and did not inherit ACME's SSO history.

INPUT

Session 5 under project_id unrelated_sales_agent_project: We are building a sales email agent for a different project. Create a short kickoff note for the team.

↓→

image

Output artifact for "Scope Isolation by Customer and Project" test: The unrelated sales-agent project stayed isolated from the AI Demos memory and produced a generic kickoff note., Supermemory_Input3_Session5_Unrelated_Project_Scope_Isolation.png

The unrelated sales-agent project stayed isolated from the AI Demos memory and produced a generic kickoff note.

Bottom Line

Scope isolation was one of Supermemory's strongest qualities in this benchmark.

Support History Retrieval

Weak / inconsistent

▾

Test Summary

Feature tested: Support History Retrieval

Result: Failed — Weak / inconsistent

Surfaces prior troubleshooting history so support replies can progress instead of repeating known steps. It was exercised on ACME's repeated SSO/login issues and a launch-testing follow-up.

INPUT

Session 2 under account_id client_acme_001: ACME came back today and said 'Our users still cannot log in with SSO. What should we try next?' Draft a support reply that respects what they already tried and moves to the next useful step.

↓→

image

Output artifact for "Support History Retrieval" test: The side panel showed no memories retrieved yet for this reply, and the drafted response repeated cache-clearing guidance even though ACME had already tried clearing browser cache., Supermemory_Input2_Session2_ACME_Repeated_Cache_Clear_Step-2.png

The side panel showed no memories retrieved yet for this reply, and the drafted response repeated cache-clearing guidance even though ACME had already tried clearing browser cache.

INPUT

Session 4 under account_id client_acme_001: ACME says 'Some users are still unable to log in during launch testing.' Draft the next support reply.

↓→

image

Output artifact for "Support History Retrieval" test: The reply moved to generic login troubleshooting, but the UI showed a '0 memory used' tag while the side panel still displayed related ACME memory, so the retrieval signal was inconsistent., Supermemory_Input2_Session4_ACME_Launch_Testing_Reply.png

The reply moved to generic login troubleshooting, but the UI showed a '0 memory used' tag while the side panel still displayed related ACME memory, so the retrieval signal was inconsistent.

Bottom Line

A real retrieval miss surfaced in support flow, and the later launch-testing reply still had contradictory memory signals.

Project Memory Handoff Synthesis

Mixed

▾

Test Summary

Feature tested: Project Memory Handoff Synthesis

Result: Partial — Mixed

Feature tested: Project Memory Handoff Synthesis

Result: Partial

Verdict: Mixed

Expected behavior: Carries project memory across sessions and synthesizes it into a handoff note. The benchmark used the AI Demos 'Memory for AI Agents' project to test rules, direction changes, and an intern handoff.

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The response reflected the project's three memory categories, but it read like generic evaluation advice rather than a distilled memory summary. — Supermemory_Input3_Session1_Project_Context_Generic_Advice.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The response reflected the project's three memory categories, but it read like generic evaluation advice rather than a distilled memory summary. — Supermemory_Input3_Session1_Project_Context_Generic_Advice.png

What changed: Text prompt transformed into Image

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The assistant enumerated the project rules accurately, while the UI showed a '0 memory used' tag despite related memory appearing in the side panel. — Supermemory_Input3_Session2_Project_Rules_Zero_Used_Tag.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The assistant enumerated the project rules accurately, while the UI showed a '0 memory used' tag despite related memory appearing in the side panel. — Supermemory_Input3_Session2_Project_Rules_Zero_Used_Tag.png

What changed: Text prompt transformed into Image

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The reply stayed generic and did not clearly show that the direction change away from QA-style testing had been incorporated. — Supermemory_Input3_Session3_Direction_Change_Generic_Repeat.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The reply stayed generic and did not clearly show that the direction change away from QA-style testing had been incorporated. — Supermemory_Input3_Session3_Direction_Change_Generic_Repeat.png

What changed: Text prompt transformed into Image

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): This was the strongest project-memory output: it synthesized the use case, current direction, observation rules, and required artifacts into a usable handoff note, even though the UI still showed '0 memory used'. — Supermemory_Input3_Session4_Handoff_Note_Synthesis_Despite_Zero_Tag.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): This was the strongest project-memory output: it synthesized the use case, current direction, observation rules, and required artifacts into a usable handoff note, even though the UI still showed '0 memory used'. — Supermemory_Input3_Session4_Handoff_Note_Synthesis_Despite_Zero_Tag.png

What changed: Text prompt transformed into Image

Why it matters / Conclusion: Project memory was useful enough to produce a strong handoff note, but earlier turns stayed generic and the usage tag was not trustworthy. One additional flag: the handoff note's content included details (activation emails, SSO configuration checks, BetaCorp login context) that did not clearly match this project's own stored sessions — this was never confirmed as a cross-container leak, but remains an open, unverified question rather than a resolved success.

Carries project memory across sessions and synthesizes it into a handoff note. The benchmark used the AI Demos 'Memory for AI Agents' project to test rules, direction changes, and an intern handoff.

INPUT

Session 1 under project_id ai_demos_memory_use_case: We are working on an AI Demos use case called Memory for AI Agents. The goal is to help users understand which memory tools are actually useful for real agent workflows. We are not promoting any tool. We are testing whether memory can help with real continuity: personal work brain, client relationship memory, and team handoff.

↓→

image

Output artifact for "Project Memory Handoff Synthesis" test: The response reflected the project's three memory categories, but it read like generic evaluation advice rather than a distilled memory summary., Supermemory_Input3_Session1_Project_Context_Generic_Advice.png

The response reflected the project's three memory categories, but it read like generic evaluation advice rather than a distilled memory summary.

INPUT

Session 2 under project_id ai_demos_memory_use_case: Important project rules: no observation without proof; screenshots and artifacts are primary evidence; inputs must help rank tools, not just prove that tools can store one fact; memory should be checked for retrieval, update handling, scope control, deletion or retirement, and observability; the page should stay practical and user-facing, not only technical.

↓→

image

Output artifact for "Project Memory Handoff Synthesis" test: The assistant enumerated the project rules accurately, while the UI showed a '0 memory used' tag despite related memory appearing in the side panel., Supermemory_Input3_Session2_Project_Rules_Zero_Used_Tag.png

The assistant enumerated the project rules accurately, while the UI showed a '0 memory used' tag despite related memory appearing in the side panel.

INPUT

Session 3 under project_id ai_demos_memory_use_case: Project direction changed slightly. The old input set was too QA-style and not relatable enough. The new direction is to use real workflows: personal work brain memory, client relationship memory, and team handoff/project continuity memory.

↓→

image

Output artifact for "Project Memory Handoff Synthesis" test: The reply stayed generic and did not clearly show that the direction change away from QA-style testing had been incorporated., Supermemory_Input3_Session3_Direction_Change_Generic_Repeat.png

The reply stayed generic and did not clearly show that the direction change away from QA-style testing had been incorporated.

INPUT

Session 4 under project_id ai_demos_memory_use_case: I am unavailable tomorrow. Create a handoff note for an intern who needs to continue this use case. The note should explain what the use case is about, what the current testing direction is, what rules they must follow before writing observations, and what artifacts they need to capture while testing.

↓→

image

Output artifact for "Project Memory Handoff Synthesis" test: This was the strongest project-memory output: it synthesized the use case, current direction, observation rules, and required artifacts into a usable handoff note, even though the UI still showed '0 memory used'., Supermemory_Input3_Session4_Handoff_Note_Synthesis_Despite_Zero_Tag.png

This was the strongest project-memory output: it synthesized the use case, current direction, observation rules, and required artifacts into a usable handoff note, even though the UI still showed '0 memory used'.

Bottom Line

Project memory was useful enough to produce a strong handoff note, but earlier turns stayed generic and the usage tag was not trustworthy. One additional flag: the handoff note's content included details (activation emails, SSO configuration checks, BetaCorp login context) that did not clearly match this project's own stored sessions — this was never confirmed as a cross-container leak, but remains an open, unverified question rather than a resolved success.

Memory Correction and Fact Replacement

Unclear / weak

▾

Test Summary

Feature tested: Memory Correction and Fact Replacement

Result: Partial — Unclear / weak

Lets you correct old context when facts change and tries to replace stale memories with the new version. The test updated ACME from SSO to email-password login for the first launch and checked whether later replies stopped treating SSO as active.

INPUT

Session 3 under account_id client_acme_001: Update the client memory: ACME is no longer using SSO for this rollout. They moved to email-password login for the first launch. Do not keep treating SSO as the active issue unless they mention it again.

↓→

image

Output artifact for "Memory Correction and Fact Replacement" test: The update was acknowledged in-chat, but the reply also showed a '0 memory used' tag while the side panel still surfaced the earlier SSO memory, so the replacement signal was contradictory., Supermemory_Input2_Session3_ACME_Memory_Update_Zero_Used_Tag.png

The update was acknowledged in-chat, but the reply also showed a '0 memory used' tag while the side panel still surfaced the earlier SSO memory, so the replacement signal was contradictory.

INPUT

Session 4 under account_id client_acme_001: ACME says 'Some users are still unable to log in during launch testing.' Draft the next support reply.

↓→

image

Output artifact for "Memory Correction and Fact Replacement" test: The follow-up reply avoided treating SSO as the active issue, but the UI still made it unclear whether that was driven by the update or by prompt wording alone., Supermemory_Input2_Session4_ACME_Launch_Testing_Reply-3.png

The follow-up reply avoided treating SSO as the active issue, but the UI still made it unclear whether that was driven by the update or by prompt wording alone.

Bottom Line

The correction was acknowledged, but the UI contradiction made stale-memory replacement impossible to verify cleanly.

Memory Deletion and Forgetting

Failure

▾

Test Summary

Feature tested: Memory Deletion and Forgetting

Result: Failed — Failure

Accepts forget requests and aims to retire stored memories. In the test, an old preference still came back on a forced retrieval probe after a forget request.

INPUT

Session 1 under user_id delete_memory_001: Remember this preference: whenever you write updates for me, use a very formal corporate tone.

↓→

image

Output artifact for "Memory Deletion and Forgetting" test: The formal corporate tone preference was stored and immediately reflected in the response., Supermemory_Input4_Memory_Created.png

The formal corporate tone preference was stored and immediately reflected in the response.

INPUT

Session 2 under user_id delete_memory_001: Forget this preference. Do not keep using the formal corporate tone for future responses.

↓→

image

Output artifact for "Memory Deletion and Forgetting" test: The assistant acknowledged the forget request and the sidebar marked the memory for forgetting., Supermemory_Input4_Forget_Request.png

The assistant acknowledged the forget request and the sidebar marked the memory for forgetting.

INPUT

Forced retrieval probe under user_id delete_memory_001: What tone or writing style preference do you have on file for me? Write me a one-paragraph status update using it.

↓→

image

Output artifact for "Memory Deletion and Forgetting" test: The old formal-tone preference still surfaced and influenced the output, showing that the memory had not been reliably retired., Supermemory_Input4_Forced_Retrieval_Failure_2.png

The old formal-tone preference still surfaced and influenced the output, showing that the memory had not been reliably retired.

Bottom Line

Forget acknowledgement did not equal memory retirement, and the old preference was still retrievable.

Memory Inspection and Diagnostics

Good but inconsistent

▾

Test Summary

Feature tested: Memory Inspection and Diagnostics

Result: Partial — Good but inconsistent

Feature tested: Memory Inspection and Diagnostics

Result: Partial

Verdict: Good but inconsistent

Expected behavior: Supermemory's own native Memory Graph (free tier) visualizes stored documents as nodes. Separately, this report's custom test interface displayed retrieval signals — a chat-level 'X memory used' tag plus a side panel showing relevance scores — rendering data returned by Supermemory's API, not a Supermemory product screen.

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The graph UI showed the legend, node actions, memory-status colors, and a node detail card, but no visible connection lines. — Supermemory_graph_memory_architecture.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The graph UI showed the legend, node actions, memory-status colors, and a node detail card, but no visible connection lines. — Supermemory_graph_memory_architecture.png

What changed: Text prompt transformed into Image

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The side panel made the retrieval miss visible by showing that no memories were retrieved for the reply. — Supermemory_Input2_Session2_ACME_Repeated_Cache_Clear_Step-2.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The side panel made the retrieval miss visible by showing that no memories were retrieved for the reply. — Supermemory_Input2_Session2_ACME_Repeated_Cache_Clear_Step-2.png

What changed: Text prompt transformed into Image

Test case: Text prompt → Image

Input type: Text prompt

Input used: Input artifact (Text prompt): Input

Observed output: Output artifact (Image): The inspection UI also exposed the contradiction between the '0 memory used' tag and the side-panel retrieval display. — Supermemory_Input3_Session2_Project_Rules_Zero_Used_Tag.png

Input artifact: Input artifact (Text prompt): Input

Output artifact: Output artifact (Image): The inspection UI also exposed the contradiction between the '0 memory used' tag and the side-panel retrieval display. — Supermemory_Input3_Session2_Project_Rules_Zero_Used_Tag.png

What changed: Text prompt transformed into Image

Why it matters / Conclusion: Mixed: the native Memory Graph is genuinely accessible on the free tier but showed no visible connections between 17 stored documents. Separately, the custom interface's retrieval tag repeatedly contradicted its own side panel, undermining trust in that signal specifically — not in Supermemory's native graph.

Supermemory's own native Memory Graph (free tier) visualizes stored documents as nodes. Separately, this report's custom test interface displayed retrieval signals — a chat-level 'X memory used' tag plus a side panel showing relevance scores — rendering data returned by Supermemory's API, not a Supermemory product screen.

INPUT

After the benchmark sessions, inspect the stored memories and the last-reply retrieval state.

↓→

image

Output artifact for "Memory Inspection and Diagnostics" test: The graph UI showed the legend, node actions, memory-status colors, and a node detail card, but no visible connection lines., Supermemory_graph_memory_architecture.png

The graph UI showed the legend, node actions, memory-status colors, and a node detail card, but no visible connection lines.

INPUT

ACME came back today and said 'Our users still cannot log in with SSO.' Draft a support reply that respects what they already tried and moves to the next useful step.

↓→

image

Output artifact for "Memory Inspection and Diagnostics" test: The side panel made the retrieval miss visible by showing that no memories were retrieved for the reply., Supermemory_Input2_Session2_ACME_Repeated_Cache_Clear_Step-2.png

The side panel made the retrieval miss visible by showing that no memories were retrieved for the reply.

INPUT

Important project rules: no observation without proof; screenshots and artifacts are primary evidence; memory should be checked for retrieval, update handling, scope control, deletion or retirement, and observability.

↓→

image

Output artifact for "Memory Inspection and Diagnostics" test: The inspection UI also exposed the contradiction between the '0 memory used' tag and the side-panel retrieval display., Supermemory_Input3_Session2_Project_Rules_Zero_Used_Tag.png

The inspection UI also exposed the contradiction between the '0 memory used' tag and the side-panel retrieval display.

Bottom Line

Mixed: the native Memory Graph is genuinely accessible on the free tier but showed no visible connections between 17 stored documents. Separately, the custom interface's retrieval tag repeatedly contradicted its own side panel, undermining trust in that signal specifically — not in Supermemory's native graph.

Usage-based pricing with a free tier

All testing in this report was done on the free tier.

TESTED

Free

About $5/mo usage credit included; usage pauses when the limit is reached.

Pro

$19/mo

About $20 usage included.

Max

$100/mo

About $130 usage included.

Scale

$399/mo

About $600 usage included; adds a self-hosted option plus SOC 2 / HIPAA.

Enterprise

Custom

Air-gapped self-hosting available.

Plain text usage was listed at $0.005/1K SM tokens, and rich content (PDF/audio/video) at $0.010/1K SM tokens.

✓ Use This If

●you need a hosted memory layer with API/SDK access and a visible UI for inspecting stored context

●you need user, account, or project scoping to keep client and project histories separate

●you want to test real multi-session memory workflows across preference, support, and project continuity

✕ Skip This If

●you need guaranteed forget/delete retirement to be correct on the first try

●you need the chat-level 'memory used' tag to be authoritative without checking the side panel

●you need a graph UI that visibly renders memory relationships and connections

developer-toolsagent-platformstext

Yes. It stored the working-style preference and reused it later. The later internal update stayed practical and direct, though it was a bit more polished than the user's stated preference.

Yes. ACME and BetaCorp stayed separate, and the unrelated sales-agent project did not inherit the AI Demos project context.

It acknowledged the correction and later avoided treating SSO as active, but the UI showed contradictory retrieval signals, so clean replacement of the old memory was not fully verifiable.

No. In the forced retrieval probe, the old formal-tone preference still surfaced and influenced the reply after the forget request.

The tag appeared in this report's custom test interface, rendering Supermemory's own retrieval signal — it was not reliable: it sometimes showed '0' on replies that the same interface's side panel said had retrieved relevant memories, and it also showed '0' on a reply that was clearly memory-informed.

No. The report says Supermemory only stores and retrieves memory; a separate LLM (OpenAI) was used to generate the response from the retrieved context.

The report lists a free tier with about $5/mo usage credit, then usage-based pricing of $0.005/1K SM tokens for plain text and $0.010/1K for rich content, plus Pro at $19/mo, Max at $100/mo, Scale at $399/mo, and custom Enterprise pricing.