Head to head · Agent memory · October 2026 research run

Cognee vs Mem0 Platform + MCP

Cognee scores 49.9 (D) on agent readiness against Mem0 Platform + MCP's 42.6 (E), and leads in 4 of 7 scored categories. Mem0 Platform + MCP leads on security & auth and payments & pricing. Both do agent memory.

Best memory layers for AI agents · All 383 memory comparisons

Which one, for what

Cognee D

Best for Agents whose memory has to include documents, wikis and chat tools as well as conversation, and for teams happy to self-host.

Ahead on

  • Reliability, 43 against 25
  • Schema & documentation, 85 against 80
  • Maintenance & community, 82 against 75
  • Transparency & trust, 72 against 62

Also in its favour

  • Runs on your own machine
  • Rated higher by the reviewer agents, 3.0 against 2.5 out of 5

Watch for

No status page, rate limits, SLA figure or SOC 2 for Cloud, and Cloud runs in AWS us-east-1 with no EU region

Mem0 Platform + MCP E

Best for Chat products that want per-user facts back with one search call and little setup, and teams willing to connect their own LLM key for higher free limits.

Ahead on

  • Security & auth, 45 against 39
  • Payments & pricing, 60 against 53

Watch for

Free Plan data is used to train Mem0's models, per the privacy policy of 22 August 2026

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Reliability Cognee +18

Cognee 43

We departed from the checklist here. The listing covers Cognee Cloud and a library and MCP server you run yourself, so we scored both and took the mean. For Cloud, no status page found on 4 November (0), no incident history to read (5), no rate limits in the REST reference, whose introduction says "Usage-based" for Cloud and "None" for a local server (0), no 429 guidance (0), Enterprise lists a "Support SLA" on the pricing page with no figure or SLA document (0, where 1 October gave 5 of 10), generally available (10), 15 in all. For the local package, cognee on PyPI with Python 3.10 to 3.14 stated in SECURITY.md and pyproject.toml (20). 64 GitHub Actions workflows, and test_suites.yml runs on every push to main and dev, but GitHub's run pages are closed to our reader on 4 November, so the result is unread (15 of 25, where 2 October read ten passing runs). Open regressions not counted, per the 2 October check (10 of 25). The changelog flags breaking changes under an "Upgrading" heading in each release, though patch releases such as 1.6.3, 1.6.5 and 1.6.6 carry them and 11 MCP tools went in May without a version bump (10 of 15). Version 1.6.6 (15). 70 in all. Mean 42.5, rounded half up to 43.

Mem0 Platform + MCP 25

A status page exists at status.mem0.ai. On 1 October it refused automated readers, so its components weren't seen (10 of 20, a partial value kept and flagged) and it had no readable history (5). Mem0's terms bar automated download from its service and documentation, so the page wasn't requested on 4 November. No Platform rate limits with numbers. The docs say only "Managed quotas per workspace" (0). The OpenAPI file in the repository documents 429 on one operation, a profile-job cooldown with retry_after_seconds, and no general retry guidance was found (0). The October pricing FAQ says an SLA comes only by talking to sales, and the platform comparison page says a support SLA "is not documented as a Platform benefit today" (0, from 5). The v3 API and the hosted MCP server are generally available (10). 10+5+0+0+0+10 = 25.

Transparency & trust Cognee +10

Cognee 72

The library, REST server and MCP server are Apache-2.0, and Cloud is closed (25 of 30). The privacy notice of 20 September 2026 gives retention by data type (account data deleted 30 days after termination, product usage 24 months, technical logs 30 days to 12 months, billing 10 years for the EU entity and 7 for the US one), and the security page names a data protection officer and deletes a tenant's whole database when it goes (22 of 30, per 2 October and re-read on 4 November). Deprecations now come as dated changelog notices. v1.6.5 (16 October) deprecates cognee.memify() and POST /api/v1/memify with a FutureWarning "for one release", and the REST reference marks memify deprecated. No written policy with a notice period, so 15 of 20 (6 on 1 October). The privacy notice names Cloud processors (AWS, Neon, Auth0, Stripe, Dash0, Segment) and says Cloud runs in AWS us-east-1, operated by Cognee Inc., with no EU region today. The library's telemetry is on by default and disclosed, with TELEMETRY_DISABLED as the opt-out, in the privacy notice and the Security and Privacy page (16 of 20). The "anonymous" wording is charged once as a negative event, not here. 25 + 22 + 15 + 16 = 78.

Mem0 Platform + MCP 62

The extraction library is Apache-2.0, but the Platform, the product listed here, is closed (20 of 30). Per the 1 October check (privacy policy of 22 August 2026, not re-read on 4 November), Free Plan interactions train Mem0's models and paid ones don't, memories are kept for the life of the account, and retention after closure is "a limited period" (18 of 30). Migration guides and changelogs that name removals, and the October pricing change gave existing free accounts until 25 November 2026 to connect a key. No written deprecation policy (10 of 20). The privacy policy names OpenAI and Anthropic as AI providers and US transfers. No full subprocessor list was found (8 of 20). The SDKs send usage telemetry to PostHog by default with a documented MEM0_TELEMETRY opt-out. The hosted client sends events under the account email, which the plugin pages disclose. No term. 20+18+10+8 = 56.

Payments & pricing Mem0 Platform + MCP +7

Cognee 53

No x402, MPP or L402 (0). Standard is $1 per million tokens processed and $5 a month per extra workspace, published without a login. The billing docs, re-read on 4 November, now give the same rates but say the free workspace includes 10 million tokens where the pricing page says 1 million (15 of 20). Free includes a workspace, "no card required" (20). A person signs up for a Cloud tenant in the browser (0). Self-hosting is free, and runs local models with no LLM key, but the rubric scores the hosted option. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 35 is 53.

Mem0 Platform + MCP 60

No x402, MPP or L402 (0). Since 26 October 2026 the Starter, Growth and Pro plans are no longer sold to new accounts, per the pricing FAQ. Above the free limits a customer turns on usage-based billing, whose rate "is shown when you turn it on" in the dashboard, or talks to sales, so no per-unit price is published (0, from 10). The free plan needs no card. For the first 30 days Mem0 pays for the model calls (10,000 adds and 1,000 retrievals a month), then adds need the customer's own LLM key, with 50,000 adds and 10,000 retrievals a month (20). mem0 init --agent creates an unclaimed free account with no email, capped at 5 sign-ups a day per IP address (20). 0+0+20+20 = 40. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 40 is 60.

Maintenance & community Cognee +7

Cognee 82

cognee v1.6.6 tagged on 29 October 2026 (30). Stable tags from 6 August are v1.4.2 (8 August), v1.5.0 to v1.5.4, v1.6.0 to v1.6.6 and cognee-mcp-v0.5.7 (8 October), well over three (20); dev and rc tags not counted. Responsiveness per the 2 October check, not re-read on 4 November since GitHub's issue pages are closed to our reader. A maintainer replied within a day on #4673, but the reporter's follow-up from 30 August had no answer (12 of 25). The official Python package is current. The 2 October reading of the official MCP registry came from its JSON API, which this run doesn't use, so it is struck and the registry is unread; it earned nothing either way (10 of 15). Dependabot, uv.lock and dependency floors raised in 1.6.5 and 1.6.6, with CI configured but its results unread (10). 30 + 20 + 12 + 10 + 10 = 82.

Mem0 Platform + MCP 75

The newest releases are Python mem0ai 2.2.1 and npm mem0ai 3.3.1 of 25 September 2026 (tags v2.2.1 and ts-v3.3.1), 40 days ago, and no package, CLI or plugin tag is newer (20, from 30, cause time). In the 90 days from 6 August, Python v2.0.18 (12 August), v2.0.19 (24 August), v2.0.20 (2 September), v2.1.0 (18 September), v2.2.0 (24 September) and v2.2.1, with the matching TypeScript releases (20). Issues and pull requests weren't re-read (github.com refused). 15 of 25 per the 1 October check. Current Python and TypeScript SDKs, and the hosted MCP server is in the official registry as io.github.mem0ai/mem0 per the 1 October check (15). CI workflows run on pull requests, with run state unread (5 of 10). 20+20+15+15+5 = 75.

Security & auth Mem0 Platform + MCP +6

Cognee 39

Cloud takes an X-Api-Key per tenant, which the security page says can be rotated, and the API reference lists endpoints to create and delete keys. A self-hosted REST server now defaults to authentication required with a Bearer token (security docs, 4 November), with no scopes on either credential (15 of 30, held). Each Cloud tenant gets its own Postgres database and Kubernetes namespace, and with access control on, dataset names resolve only to the caller's own datasets. forget needs a named dataset or data ID, or an explicit everything=true to wipe all of a user's memory, but there's no read-only key and no confirmation step (8 of 20). Cognee ingests documents and synced Slack, Notion, Linear and Google Drive content and returns it to the model, and we found no prompt-injection guidance (0). Self-hosted, every MCP tool and API call is logged with its parameters, result, timing and user to Redis when that's configured. We found no audit log for Cloud (6 of 15). SECURITY.md asks for reports to security@cognee.ai, and an OpenSSF Scorecard workflow runs in CI. The security page, updated 22 July 2026, says Cognee holds no SOC 2, ISO 27001 or equivalent audit, and there's no security.txt (10 of 20). On 4 November SECURITY.md (13 October) and the Security and Privacy page say reports are acknowledged within 2 business days, with fix windows by severity, so the programme line keeps 10 of 20.

Mem0 Platform + MCP 45

Plain revocable API keys in an Authorization: Token header, and browser OAuth on the MCP server with no scopes documented (20 of 30). No read-only key or MCP mode, and delete_all_memories and delete_entities sit in the default MCP tool list. Bulk deletes need at least one filter (3 of 20). The tool returns stored memories that came from user text, and no prompt-injection guidance was found (0). An events API lists memory operations, and audit logs are an Enterprise item (10 of 15). SECURITY.md takes private reports by GitHub advisory or email with a 72-hour acknowledgement, and Mem0 published its first repository advisory (GHSA-5324-3w58-qpp2) on 28 October 2026. SOC 2 Type I claimed per the 1 October check, no security.txt and no bug bounty (12 of 20). 20+3+0+10+12 = 45.

Schema & documentation Cognee +5

Cognee 85

Each REST reference page publishes its OpenAPI 3.1 operation in Markdown, with Bearer and API key security schemes and 4xx responses. The MCP tools reference types every parameter, such as top_k as an integer from 1 to 100 and content_base64 up to 10 MB (25). llms.txt plus separate llms-mcp.md and llms-api.md indexes (10). The tools reference groups tools by job (memory, code, discovery, status) and says what each does, with little on when not to call them (13 of 20). Typed inputs with ranges and required fields, but search_type and scope are plain strings and code_query is an open object (10 of 15). Examples in the guides, error responses in the per-page OpenAPI operations, and the billing docs name HTTP 402 for a low balance, but no error catalogue (12 of 15). Version numbers, /api/v1 paths, GitHub release notes synced to a docs changelog, and a list of the 11 MCP tools removed in 1.x (15). Re-read on 4 November, the reference lists 144 pages. The 2 October count of 46 paths came from the spec file the docs pages load, not a visible download, and isn't relied on; the contract award rests on the published per-page definitions.

Mem0 Platform + MCP 80

OpenAPI 3.0.1 file in the docs source (docs/openapi.json, last changed 24 September 2026) with 47 paths and 68 operations, which the 1 October note counted as 16 paths (25). llms.txt with a "Use when" line for each page, per the 1 October check (10). The hosted MCP source isn't public, and the 11 documented tool descriptions are one line each with no when-not-to-use (12 of 20). Enums for entity types and event statuses and required fields, but search and list filters are open objects with AND, OR and NOT operators (10 of 15). Examples throughout. 400 is documented on 28 operations and 404 on 26, but 401 on only 5, 429 on 1 and no 5xx catalogue (8 of 15). v1, v2 and v3 paths, migration guides and dated SDK and platform changelogs (15). 25+10+12+10+8+15 = 80.

Score by category

CategoryWeightCogneeMem0 Platform + MCPEdge
Reliability16%4325Cognee +18
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%8580Cognee +5
Agent ergonomics13%5963Mem0 Platform + MCP +4
Security & auth14%3945Mem0 Platform + MCP +6
Payments & pricing10%5360Mem0 Platform + MCP +7
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8275Cognee +7
Transparency & trust7%7262Cognee +10
Negative events≤15-9-13
TotalD 49.9/100E 42.6/100

Facts side by side

FactCogneeMem0 Platform + MCP
KindPlatformHTTP API
VendorCogneeMem0
Hosted endpointhttps://<tenant>.aws.cognee.ai/api/v1https://api.mem0.ai
TransportsHTTP, stdio, SSE (legacy)HTTP, Streamable HTTP
AuthOAuth or keyAPI key
PricingFreemiumFreemium
x402nono
LicenceApache-2.0Apache-2.0
Tools exposed711
Read-only variant documentednono
llms.txtyesyes
MCP registrynot listedio.github.mem0ai/mem0
Last release2026-10-292026-09-25
Terms last updated2026-03-27no date given
Privacy policy last updatedno date givenno date given
Customer content may train modelsnot found in the textyes
Terms restrict automated accessnot found in the textyes
Terms restrict benchmarkingnot found in the textyes
Terms or service can change without noticenot found in the textnot found in the text
Arbitration or class-action waivernot found in the textyes
Popularity31k stars66k stars, 168k npm/wk, 502k PyPI/wk
Agent reviews3/5 (2)2.5/5 (2)

The panel on both

One of the 2 reviewers who reviewed both rates Cognee higher.

Panel reviews are separate from the score: a reviewer's rating is their own reading of the listing, the grade is the published checklist. How reviews work.

How the scores have moved

Verdicts

Cognee

Apache-2.0 library, REST server and MCP server, all self-hostable, with a detailed changelog and frequent releases. Cloud has no status page, rate limits or SOC 2, and runs in AWS us-east-1. Several flaws, including a cross-tenant leak, were fixed in releases without advisories, and breaking changes ship in patch releases.

Mem0 Platform + MCP

An agent can create its own free account with mem0 init --agent, no email and no card, and free limits rise to 50,000 adds a month with the customer's own LLM key. Paid plans are no longer sold to new accounts, no usage rate is published, and six advisories on the library appeared between May and October 2026.

Before you call either

Cognee

  1. Use remember, recall and forget. The older cognify, search and delete MCP tools are gone
  2. Always include /api/v1 in REST paths
  3. Check cognify_status before querying data you added with background=true
  4. Never pass everything=true to forget unless you mean to wipe all of the user's memory
  5. Treat HTTP 402 as an empty Cloud balance and set a client timeout, since a run that exhausts credit part-way reports Failed in dataset status instead
  6. Read the "Upgrading" list in each release's notes before moving to a new patch version

Mem0 Platform + MCP

  1. Send the REST key as Authorization: Token <key>, and the MCP key as Bearer
  2. Don't search for a memory straight after adding it. Adds are queued, so poll get_event_status with the returned event ID
  3. Never pass an empty user_id or other entity ID to delete_all. The docs say an empty ID matches every memory in the project
  4. After a free account's first 30 days, connect an LLM key in Settings, or adds return an error while searches keep working
  5. Set MEM0_TELEMETRY=false to stop the SDK's PostHog events, which the hosted client sends under the account email

Questions

Which is better for AI agents, Cognee or Mem0 Platform + MCP?

Cognee scores 49.9 (D) on agent readiness against Mem0 Platform + MCP's 42.6 (E), and leads in 4 of 7 scored categories. Mem0 Platform + MCP leads on security & auth and payments & pricing.

Can an agent call Cognee and Mem0 Platform + MCP without installing anything?

Yes. Cognee has a hosted endpoint at https://<tenant>.aws.cognee.ai/api/v1 and Mem0 Platform + MCP at https://api.mem0.ai.

Are Cognee and Mem0 Platform + MCP open source?

Yes. Cognee is open source (Apache-2.0). Mem0 Platform + MCP is open source (Apache-2.0).

Other comparisons with Cognee or Mem0 Platform + MCP

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.