AI memory comparison
Tephra vs Cognee
Cognee is an Apache-2.0 AI memory platform with a free-forever plan, token-based pricing and connectors that read Slack, GitHub and Google Drive into a shared workspace memory. Tephra is a hosted memory that the AI apps you connect share, set up by signing in. Cognee is the better fit if you want a company brain fed from your existing tools, or you want to run it yourself.
Sources read and prices checked 2026-10-10.
Obsidian Metrics makes Tephra and Tracker, the two products this page compares against others. This page is written by the maker of one of the products compared, so read the competitor columns against their own pages, linked below each table. Competitor facts come only from the vendor pages listed under Sources, read on the date shown above; vendors change plans and prices without notice. We are not affiliated with any company named here, and all product names belong to their owners.
Tephra and Cognee side by side
| Feature | Tephra | Cognee |
|---|---|---|
| What it is | A hosted memory your connected AI apps share: they recall from it and write to it over MCP or REST. | In its own words, "the open-source AI memory platform for agents". Its MCP server gives assistants such as Claude Code, Cursor and Codex persistent memory backed by a knowledge graph. [1] [2] |
| Entry price | Tephra needs a paid plan: Premium at $19 a month in the Obsidian Metrics Skool community or $29 a month direct, or a $399 one-time seat. No free tier. | A free plan it calls "free forever": 1 workspace and 1M tokens, no card required, with unlimited users and unlimited API calls. Standard is $1.00 per 1M tokens processed plus $5 per additional workspace. Enterprise is quoted. Read 2026-10-10. [3] |
| How usage is limited | No usage metering. There are only abuse limits: 60 requests a minute per key and 120 a minute per member. | Metered on tokens processed rather than requests or seats. Seats are explicitly not metered on the free plan and API calls are explicitly unmetered; extra workspaces are a fixed $5 a month each. [3] |
| Reach across AI apps | One memory for every AI you connect, led by OAuth sign-in: the Connect page has steps for Claude (web, Desktop, mobile), ChatGPT, VS Code Copilot, Cursor, Claude Code, Gemini CLI, Windsurf, Zed, Mistral Le Chat and Perplexity, plus generic steps for any other MCP app, and a personal key as a fallback, plus REST. AI tools that support neither MCP nor REST calls cannot connect. Tephra's OAuth discovery endpoints were verified live on 2026-10-10. | An MCP server you run yourself, in Standalone Mode with its own local database or in API Mode against a shared Cognee backend, naming Claude Code, Claude Desktop, Cursor, Cline, Continue and Codex. We found no hosted remote MCP URL and no OAuth documented, so a user must run a process and supply a token; its docs state that Cognee MCP and Cognee Cloud are separate systems with different APIs and authentication schemes. [2] |
| Seeing and correcting what is stored | A member can view memories, edit or delete them, read edit history, contest a memory, export their own data (app session), and revoke any AI key with an activity log. | A web UI in its hosted cloud for exploring the knowledge graph, running searches and managing datasets, and a first-class forget operation beside remember and recall. Conflict resolution and per-answer provenance are sold as enterprise features. We did not confirm per-memory version history. [4] [3] |
| Hosting and self-hosting | Hosted service run by Obsidian Metrics. Self-hosting is not part of the product. We make no encryption-at-rest or compliance-certification claim. | Apache-2.0, and the UI is self-hostable free with no account. Two caveats it discloses itself: its Postgres graph store is released as a demo feature with the production version licensed separately, and its enterprise tier includes a proprietary runtime, so the fastest path is not the open part. Hosted workspaces sleep when idle. [1] [4] [3] |
| How memories get in | The connected AI writes them by calling a remember tool. Tephra does not capture chats automatically and has no Gmail, Slack or Drive connectors. You can upload text, markdown, JSON, XML and text-layer PDF files. | An explicit remember call or an ingestion pipeline, plus real connectors: its cloud reads Slack, GitHub and Google Drive into a shared workspace memory, and Slack, Notion, Linear and Google Drive integrations are listed on its paid tier. [4] [3] |
Where Cognee is better
- Cognee. Lets a buyer start with no card and no sales call on a free plan that is not a token: unlimited users and unlimited API calls. It is Apache-2.0 including a self-hostable UI, it prices at a flat $1.00 per 1M tokens that anyone can model on a spreadsheet, and it ingests Slack, Google Drive, GitHub, Notion and Linear - so its memory fills from where work already happens, while Tephra's fills only when an AI calls its save tool. It also sells conflict resolution and per-answer provenance, which Tephra asks you to do by hand. [3] [4] [1]
Choose Cognee if: You want to try something for free with no card, you need to self-host, or - the big one - you want memory that fills itself from Slack, Google Drive, GitHub, Notion or Linear. Tephra has no connectors of any kind, so for that job it is not in the comparison.
Where Tephra is better
- Connected AI apps read and write the same memory directly, rather than moving it between apps by pasting text. Claude's import and export, for example, works by pasted text, and its import path is marked experimental.
- A flat price with no usage metering, instead of credits or request quotas.
- A shared library that Tephra grows every hour from public research and regulatory sources, included with the product. We did not find a competitor that bundles one; that is a qualitative check, not an exhaustive market survey.
- Recall in the app tells you which related memories it did not load and how complete the answer looks.
Choose Tephra if: You want a hosted memory your MCP apps reach in a few clicks by signing in with OAuth, rather than running a process and pasting a token: we found no hosted remote MCP endpoint and no OAuth in Cognee's documentation.
Where Tephra falls short
- Tephra has no free tier and no self-hosted option.
- The MCP tools return ranked records only. The completeness alert is shown in the app, not through MCP.
- No automatic chat capture, and no connectors to Gmail, Slack or Drive.
- No image or audio ingest, and no OCR for scanned PDFs.
- Stale-fact handling is partial.
- No published, vendor-comparable benchmark, and no accuracy lead we can defend.
What we have and have not measured
On 50 held-out LongMemEval-S questions never used in Tephra development, Tephra answered 39 of 50 correctly (78%, 95% confidence interval 65-87%) with a gpt-4o-mini reader, and every evidence session was in its top 10 results for 46 of 46 answerable questions.
Method: LongMemEval-S cleaned, stratified 100-question subset; gpt-4o-mini reader over top-10 retrieved items; official GPT-4o judge prompt; one run, 2026-10-08; held-out questions are the 50 never used in earlier Tephra tuning; competitors run locally from open source under the same harness.
In that same-harness run, none of Tephra's leads over the other systems tested was statistically significant after correcting for multiple comparisons, and its margin over a plain keyword-search baseline was not significant either. We do not claim Tephra is the most accurate memory. We did not benchmark Supermemory, Zep, ChatGPT memory or Claude memory, and we do not compare ourselves with any vendor's self-reported numbers.
Frequently asked
Is Cognee free?
Cognee lists a plan it calls free forever with 1 workspace and 1M tokens and no card required, including unlimited users and unlimited API calls. Its Standard tier is $1.00 per 1M tokens processed plus $5 per additional workspace, as read on 2026-10-10. Tephra has no free tier.
Can Cognee read my Slack or Google Drive?
Yes. Cognee's cloud documentation describes reading Slack, GitHub and Google Drive into a shared workspace memory, and its paid tier lists Slack, Notion, Linear and Google Drive integrations. Tephra has no Gmail, Slack or Drive connectors and does not capture chats automatically.
Which one handles out-of-date facts better?
Cognee, on the evidence we can cite. It sells bi-temporal memory with conflict resolution and per-answer provenance on its enterprise tier. Tephra's handling of stale facts is partial: you edit or delete the old memory yourself.
Did you benchmark Tephra against Cognee?
No. We have not benchmarked Cognee and we make no accuracy claim against it.
More comparisons
Sources
Competitor facts come only from the pages below, accessed 2026-10-10.
- Cognee source repository (Apache-2.0): https://github.com/topoteretes/cognee (accessed 2026-10-10)
- Cognee documentation, MCP overview: https://docs.cognee.ai/cognee-mcp/mcp-overview (accessed 2026-10-10)
- Cognee pricing: https://cognee.ai/pricing (accessed 2026-10-10)
- Cognee documentation, Cognee Cloud overview: https://docs.cognee.ai/cognee-cloud/overview (accessed 2026-10-10)