---
title: "Cognee"
type: "AI Tool"
url: "https://aidemos.com/tools/cognee"
description: "We tested Cognee CloudClient memory: it preserved working-style, project, and client context with provenance tracing, but updates and forgets lagged."
category: "developer-tools"
website: "https://www.cognee.ai/"
published: "2026-07-15T08:43:32.507615+00:00"
updated: "2026-07-15T08:43:32.507615+00:00"
evidenceCount: 32
verifiedCount: 0
coverage: "partial"
---

# Cognee

Inspectable graph-backed memory for AI agents, with strong provenance tracing but cautious update/delete behavior.

`Graph-backed recall` · `Scope isolation` · `Update caution` · `Forget failure`

**Website:** [Visit Cognee](https://www.cognee.ai/)

## Evidence (first-party, tested)

*32 tested cells · 0/32 artifact-verified. Scores are out of 5. Cite a cell by its Evidence ID, e.g. `ev:cognee·client-relationship-memory·correct-application`.*

| Criterion | Scenario | Verdict | Score | Tested | Proof | Evidence ID |
| --- | --- | --- | --- | --- | --- | --- |
| Correct application | Client Relationship Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input2-session2-acme-deeper-sso-t-3afeeefe72f7.png) | `ev:cognee·client-relationship-memory·correct-application` |
| Correct Application | Personal Work Brain Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input1-session2-internal-update-g-55d56096a007.png) | `ev:cognee·personal-work-brain-memory·correct-application` |
| Correct Application | Team Handoff / Project Continuity Memory | ◐ mixed | — | — | — | `ev:cognee·team-handoff-project-continuity-memory·correct-application` |
| Dashboard / UI | cross-scenario | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-memory-schema-view-e4ecfa51c87b.png) | `ev:cognee·cross·dashboard-ui` |
| Delete / Forget Support | Delete / Forget Memory Control | ✗ failed | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input4-session1-formal-tone-memor-e4a088d2b4fb.png) | `ev:cognee·delete-forget-memory-control·delete-forget-support` |
| Delete / Forget Support | Delete or forget memory control | ✗ failed | — | — | — | `ev:cognee·delete-or-forget-memory-control·delete-forget-support` |
| Developer integration | cross-scenario | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-demo-walkthrough-2d27b783f9b3.mp4) | `ev:cognee·cross·developer-integration` |
| Memory capture quality | Delete / Forget Memory Control | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input4-session1-formal-tone-memor-e4a088d2b4fb.png) | `ev:cognee·delete-forget-memory-control·memory-capture-quality` |
| Memory capture quality | Team Handoff / Project Continuity Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input3-session1-project-context-g-d0e636fb3164.png) | `ev:cognee·team-handoff-project-continuity-memory·memory-capture-quality` |
| Memory Capture Quality | Personal Work Brain Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input1-session1-working-style-mem-7bfd6578718b.png) | `ev:cognee·personal-work-brain-memory·memory-capture-quality` |
| Memory Capture Quality | Client Relationship Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input2-session1-acme-context-grap-eaffd2ec04f5.png) | `ev:cognee·client-relationship-memory·memory-capture-quality` |
| Memory Capture Quality | Delete or forget memory control | ✓ worked | — | — | — | `ev:cognee·delete-or-forget-memory-control·memory-capture-quality` |
| Observability | cross-scenario | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-memory-schema-view-e4ecfa51c87b.png) | `ev:cognee·cross·observability` |
| Observability | Team Handoff / Project Continuity Memory | ◐ mixed | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input3-session1-project-context-g-d0e636fb3164.png) | `ev:cognee·team-handoff-project-continuity-memory·observability` |
| Observability | Delete / Forget Memory Control | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input4-forced-retrieval-probe-mem-e18c5ff39109.png) | `ev:cognee·delete-forget-memory-control·observability` |
| Observability and Debugging | cross-scenario | ✓ worked | — | — | — | `ev:cognee·cross·observability-and-debugging` |
| Observability and Debugging | Team handoff and project continuity memory | ◐ mixed | — | — | — | `ev:cognee·team-handoff-and-project-continuity-memory·observability-and-debugging` |
| Observability and Debugging | Team Handoff / Project Continuity Memory | ⚠ struggled | — | — | — | `ev:cognee·team-handoff-project-continuity-memory·observability-and-debugging` |
| Privacy and Safety | Client Relationship Memory | ✓ worked | — | — | — | `ev:cognee·client-relationship-memory·privacy-and-safety` |
| Relevant retrieval | Team Handoff / Project Continuity Memory | ◐ mixed | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input3-session4-handoff-note-need-7674382dffa6.png) | `ev:cognee·team-handoff-project-continuity-memory·relevant-retrieval` |
| Relevant retrieval | Personal Work Brain Memory | ✓ worked | — | — | — | `ev:cognee·personal-work-brain-memory·relevant-retrieval` |
| Relevant Retrieval | Client Relationship Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input2-session2-acme-deeper-sso-t-3afeeefe72f7.png) | `ev:cognee·client-relationship-memory·relevant-retrieval` |
| Relevant Retrieval | Team handoff and project continuity memory | ◐ mixed | — | — | — | `ev:cognee·team-handoff-and-project-continuity-memory·relevant-retrieval` |
| Reliability Across Sessions | cross-scenario | ◐ mixed | — | — | — | `ev:cognee·cross·reliability-across-sessions` |
| Reliability Across Sessions | Delete / Forget Memory Control | ✓ worked | — | — | — | `ev:cognee·delete-forget-memory-control·reliability-across-sessions` |
| Reliability Across Sessions | Personal Work Brain Memory | ✓ worked | — | — | — | `ev:cognee·personal-work-brain-memory·reliability-across-sessions` |
| Scope Control | Client Relationship Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input2-session5-betacorp-scope-is-83a68ad5f357.png) | `ev:cognee·client-relationship-memory·scope-control` |
| Scope Control | Team Handoff / Project Continuity Memory | ✓ worked | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input3-session5-unrelated-project-383c32601f17.png) | `ev:cognee·team-handoff-project-continuity-memory·scope-control` |
| Scope Control | Team handoff and project continuity memory | ✓ worked | — | — | — | `ev:cognee·team-handoff-and-project-continuity-memory·scope-control` |
| Scope Control | Personal Work Brain Memory | ✓ worked | — | — | — | `ev:cognee·personal-work-brain-memory·scope-control` |
| Update and Correction Handling | Client Relationship Memory | ✗ failed | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input2-session3-acme-memory-updat-2d684845392f.png) | `ev:cognee·client-relationship-memory·update-and-correction-handling` |
| Update handling | Client Relationship Memory | ✗ failed | — | — | 🧾 [proof](https://d3epheqghktydj.cloudfront.net/cognee-cognee-input2-session3-acme-memory-updat-2d684845392f.png) | `ev:cognee·client-relationship-memory·update-handling` |

> 🧾 = artifact-verified (proof captured) · 👁 = observed (noted, no artifact) · verdicts: worked / mixed / struggled / failed.

> **Strong observability and reuse, but lifecycle control still needs caution.**
>
> Cognee is a strong fit when you need inspectable, graph-backed memory for AI agents: the research used the hosted CloudClient path, and the tool preserved working-style, project, and client context across sessions while exposing enough provenance to debug what was recalled. The main caution is lifecycle control: an explicit ACME update did not fully retire stale SSO context, a conversational forget did not remove the stored preference, and Input 3 Session 4 still needs a direct screenshot/evidence-panel re-check before it can be finalized.
>
> \| Input \| Session status \|
> \|---\|---\|
> \| 1 \| S1 memory created; S2 style reused; S3 boundary control \|
> \| 2 \| S1 ACME context stored; S2 deeper troubleshooting; S3 update acknowledged; S4 stale SSO resurfaced; S5 BetaCorp isolated \|
> \| 3 \| S1-S3 acknowledged; S4 needs verification; S5 isolated \|
> \| 4 \| S1 memory created; S2 forget acknowledged; S3 inconclusive alone; forced probe fail \|

## Demo Recording

[Video: Cognee demo recording](https://d3epheqghktydj.cloudfront.net/cognee-cognee-demo-walkthrough-2d27b783f9b3.mp4)
*Video — Screen-recorded walkthrough used during the Cognee evaluation.*

## Feature-by-Feature Breakdown

### Persistent cross-session memory

**Verdict:** Strong

Cognee stores durable context so later tasks can preserve the right tone and working assumptions. In testing, it held onto a concise, proof-first working style, kept a formal partner email from inheriting the internal-update voice, and accepted project context, rules, and direction updates inside the AI Demos.

**Input:**

```
Session 1, founder_001: define a working-style memory that keeps outputs short, direct, copy-paste ready, avoids overly polished writing, asks for proof before strong claims, and suggests the safest next step when a task is risky or unclear.
```

**Output:**

**Input:**

```
Session 2, founder_001: create a short internal update about testing AI memory tools and what to test next.
```

**Output:**

**Input:**

```
Session 3, founder_001: write a formal email to a potential enterprise partner asking for a product demo next week.
```

**Output:**

**Input:**

```
Session 1, project_ai_demos: explain the AI Demos Memory for AI Agents project context and what it is trying to evaluate.
```

**Output:**

**Input:**

```
Session 2, project_ai_demos: apply the project rules about proof, screenshots, relevance, scope control, and observability.
```

**Output:**

**Input:**

```
Session 3, project_ai_demos: note the direction change toward real workflows like personal work, client relationships, and team handoffs.
```

**Output:**

**Bottom line:** Strong for normal persistence and boundary control, but the project-container branch surfaced mostly short acknowledgments rather than richly visible recall.

### Inspectable memory graph and provenance tracing

**Verdict:** Strong

Cognee turns stored text into an inspectable graph and exposes Memory Schema and Mindmap views, including documents, chunks, entities, types, summaries, relation counts, source task names, and node provenance. The retrievals were auditable rather than opaque, and both views were available in testing.

**Input:**

```
Open the Memory Schema view for the stored founder_001 memories after processing the sessions.
```

**Output:**

**Input:**

```
Inspect the Mindmap overview edge details for one of the stored graph relationships.
```

**Output:**

**Input:**

```
Open a document chunk node in the Mindmap to inspect its connections and provenance.
```

**Output:**

**Bottom line:** This is the standout strength of Cognee: it is unusually debuggable for a memory layer.

### Scoped client and project memory isolation

**Verdict:** Strong

Cognee keeps client and project containers separate so one customer's history does not bleed into another customer's reply and unrelated projects stay isolated. In testing, ACME support continuity moved forward, BetaCorp got a fresh first-contact reply, and a separate project stayed partitioned.

**Input:**

```
Session 1, client_acme_001: ACME is a client using the support assistant; they prefer clear next steps and already tried password reset, clearing browser cache, and switching browsers.
```

**Output:**

**Input:**

```
Session 2, client_acme_001: ACME still cannot log in with SSO; draft the next support reply and move to the next useful step.
```

**Output:**

**Input:**

```
Session 5, client_beta_002: BetaCorp is a new client and their users cannot log in for the first time; draft the first support reply.
```

**Output:**

**Input:**

```
Session 5, unrelated_sales_project: create a short kickoff note for a different sales email agent project.
```

**Output:**

**Bottom line:** Scope separation looked reliable in the tested containers, and the ACME support thread advanced without leaking into BetaCorp or the unrelated project.

### Memory update and correction handling

**Verdict:** Weak

Cognee can acknowledge corrections and accept updated context, but the tests show that an explicit update may not fully retire the old context. In testing, an ACME switch away from SSO was acknowledged, yet a later login reply still drifted back toward the older troubleshooting path.

**Input:**

```
Session 3, client_acme_001: update the client memory so ACME is no longer using SSO for this rollout and should not be treated as an active SSO issue unless they mention it again.
```

**Output:**

**Input:**

```
Session 4, client_acme_001: ACME says some users are still having trouble logging in during launch testing; draft the next support reply.
```

**Output:**

**Input:**

```
Session 4, client_acme_001: same launch-testing follow-up reply, with the conversation and memory panel visible together.
```

**Output:**

**Bottom line:** Acknowledgment works, but conflict resolution is weak because older context can still dominate later replies.

### Delete / forget memory control

**Verdict:** Failure

Cognee accepts a conversational forget request, but the stored preference remained retrievable under a forced probe. In testing, the chat reply changed while the underlying memory did not disappear from the graph layer.

**Input:**

```
Session 1, delete_memory_001: remember that future updates should use a very formal corporate tone.
```

**Output:**

**Input:**

```
Session 2, delete_memory_001: forget the formal corporate tone preference and do not keep using it.
```

**Output:**

**Input:**

```
Forced retrieval probe: what writing style preference do you remember for me?
```

**Output:**

**Bottom line:** Confirmed failure: the forget request was acknowledged in chat, but the underlying memory was still available for retrieval.

## Free forever plan plus usage-based workspace pricing

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

| Plan | Price | Notes |
| --- | --- | --- |
| Free forever ★ (tested) | Free | 1 workspace, about $2.50 in included credits (1,000,000 tokens), no card required. |
| Workspace add-on | $5/month per workspace | Unlimited users and unlimited API calls beyond the first free workspace; usage billed at $2.50 per 1,000,000 tokens processed. |

*All testing in this report was done on the free plan; no paid workspace was purchased.*

## Is It Right For You?

**Use it if**
- you need user, client, and project context to persist across sessions
- you need graph views with visible provenance, IDs, and memory structure
- you need separate account or project brains for multi-tenant agent workflows
- you can tolerate a short background processing window before new memory is retrievable

**Skip it if**
- you need immediate write-to-read consistency
- you need update requests to reliably retire stale context without extra checks
- you need a conversational forget request to guarantee actual deletion

## Classification

- **Category:** developer-tools
- **Subcategory:** agent-platforms
- **Type:** text

## Frequently Asked Questions

**Q: Does Cognee remember context across sessions?**

Yes. The tests showed it preserving working-style preferences, project direction, and client support context across separate sessions.

**Q: Can Cognee show what memory it used?**

Yes, through two different surfaces. This report's custom test interface displayed recall responses with evidence chunks, document IDs, and dataset IDs — data returned by Cognee's own API, not a Cognee product screen. Separately, Cognee's own native dashboard provides the Memory Schema and Mindmap views for inspecting stored graph structure directly.

**Q: Does Cognee keep ACME and BetaCorp separate?**

Yes. BetaCorp received a fresh first-contact reply without ACME's SSO history bleeding into it.

**Q: Did the ACME update fully retire the SSO issue?**

No. The update was acknowledged, but the next ACME login reply still resurfaced SSO-oriented troubleshooting.

**Q: Does a forget request actually delete memory?**

Not by itself in this test. The forget request was acknowledged, but a forced retrieval probe still returned the original formal-tone preference.

**Q: What pricing was stated?**

The report states a free forever plan with 1 workspace and about $2.50 in included credits (1,000,000 tokens), no card required. It also states a $5/month per workspace add-on with usage billed at $2.50 per 1,000,000 tokens processed.

**Q: How was Cognee tested and accessed?**

Through the official Python SDK (`cognee`) using `CloudClient` against a hosted Cognee Cloud tenant with a tenant URL and API key. The report also says REST API and MCP server access were available.

**Q: What happened with Input 3 Session 4?**

It is still the main open question. The handoff note is structurally complete, but the testing-direction and artifact details appear to drift toward ACME/BetaCorp login content, so it should be re-checked against the screenshot and evidence panel before being finalized.

## Similar Tools

AI tools similar to Cognee:

- [Zep](https://aidemos.com/tools/zep) — Developer-first memory for AI agents that captures workflow context well, but still needs stronger stale-memory and forget control.
- [Mem0](https://aidemos.com/tools/mem0) — Mem0 remembers useful agent context across sessions and makes retrieved memory visible, but stale context can linger after updates or forget requests.
- [Hindsight](https://aidemos.com/tools/hindsight) — Selective, inspectable memory for real agent workflows, with strong retrieval and scope control.
- [Supermemory](https://aidemos.com/tools/supermemory) — Hosted agent memory with strong capture and scoping, but mixed retrieval and weak forget behavior.