TrueMemory is live · free plan; paid plans on request · last verified 22 Oct 2026 · viability monitor
TrueMemory
TrueMemory gives CLI coding agents persistent, local-first memory recall in about 55ms — one file, zero cloud.
For a solo dev who lives in the terminal, TrueMemory is the cheapest credible fix for coding-agent amnesia: one pip install, one SQLite file, no per-query cost. The published paper and per-category LoCoMo results mean you can re-run the claims instead of trusting a vendor scoreboard. What you surrender is reach — memory stays on one machine, there's no GUI, and team sharing isn't the point.
Verdict generatedVerified 8d ago · liveness 62/100 · cite: rightaichoice.com/tools/truememory
- Solo developers running Claude Code, Cursor, Codex, or Gemini CLI who want agent memory that persists between sessions
- Privacy-conscious developers who need context to stay on the machine, with no cloud calls and no per-query bill
- Researchers who want reproducible memory benchmarks with public scripts, eval configs, and per-category results
- Developers on long-running projects where the same architectural decisions keep getting re-explained to fresh sessions
- Teams needing shared memory that syncs across devices or between teammates
- Non-technical users — setup is a terminal install with agent config edits and no GUI
- Web or mobile app developers who need a hosted memory API they can call from a product front end
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip TrueMemory if you need memory that follows you across devices or between teammates, or if you want a hosted API you can call from a web or mobile front end instead of a CLI install.
The free coding version covers individuals only — anything shipped inside a product moves to the closed-source, licensed-per-deployment tier priced per deployment rather than per seat.
The coding version of TrueMemory is free for individual developers, which undercuts hosted memory layers that charge per query or per stored memory — Mem0 and Supermemory both sit in that hosted, cross-device territory. Teams that need shared memory across devices will find TrueMemory's free tier fits a solo dev's budget but not a team's requirements, and the licensed per-deployment option is scoped case by case for ambient, legal, and healthcare products.
In short
TrueMemory — TrueMemory gives CLI coding agents persistent, local-first memory recall in about 55ms — one file, zero cloud. Best for Solo developers running Claude Code, Cursor, Codex, or Gemini CLI who want agent memory that persists between sessions, Privacy-conscious developers who need context to stay on the machine, with no cloud calls and no per-query bill, Researchers who want reproducible memory benchmarks with public scripts, eval configs, and per-category results. Free to start; it also has paid plans whose price we could not confirm — see the pricing table.
What is TrueMemory?
TrueMemory — TrueMemory gives CLI coding agents persistent, local-first memory recall in about 55ms — one file, zero cloud. Best for Solo developers running Claude Code, Cursor, Codex, or Gemini CLI who want agent memory that persists between sessions, Privacy-conscious developers who need context to stay on the machine, with no cloud calls and no per-query bill, Researchers who want reproducible memory benchmarks with public scripts, eval configs, and per-category results. Free to start; it also has paid plans whose price we could not confirm — see the pricing table.
TrueMemory is a local-first memory layer built for AI coding agents running in the terminal. It watches what your agent does, scores each moment for surprise, and writes only the roughly 7.2% it deems worth keeping — the vendor frames that as attention rather than storage. Everything lands in a single SQLite file at ~/.truememory/memory.tm on your own machine, so there are no cloud calls and no per-query bill. The engine is grounded in a published paper, arXiv:2605.04897 — "Storage Is Not Memory: A Retrieval-Centered Architecture for Agent Recall" — and runs as six layers of code. A hippocampus-style surprise gate writes only what is unexpected, a salience mechanism strengthens memories through reuse while letting stale ones fade, a consolidation pass replays recent activity to keep the signal, and a contradiction layer reconciles conflicting beliefs into one. The vendor reports 93.0% LoCoMo accuracy, 96.5% multi-hop reasoning, and ~55ms recall on a 30MB base footprint. Onboarding is a single pip install that also attaches the MCP server and rewrites configs for the CLI agents it detects — Claude Code, Cursor, Codex, and Gemini CLI. The coding version is free for individual developers, and the vendor reports 10,000+ installs and 363 GitHub stars. Anything shipped inside a product — ambient AI, legal intelligence, and healthcare are the deployments the vendor names — is closed source and licensed per deployment, with the vendor citing about 30 minutes to scope yours. Positioning matters here: TrueMemory deliberately avoids the hosted, cross-device, team-shared lane that Mem0 and Supermemory occupy. It keeps memory on one machine and lets you audit the numbers yourself, since the paper, benchmark scripts, eval configs, and per-category results are all public rather than locked behind a dashboard.
Key features
- Automatic memory capture and recall for CLI coding agents
- Surprise gate writes only unexpected information, retaining ~7.2% of observed input
- Salience weighting strengthens frequently used memories and fades unused ones
- Consolidation pass replays recent activity and keeps the signal
- Contradiction detection reconciles conflicting beliefs into one
- Six-layer cognitive architecture modeled on neuroscience mechanisms
- Multi-hop reasoning reported at 96.5% on vendor benchmarks
- Temporal reasoning across sessions
- 93.0% reported LoCoMo accuracy
- Single SQLite file storage at ~/.truememory/memory.tm
- 100% local, offline execution with zero cloud calls
- ~55ms recall latency
- 30MB base footprint
- MCP server integration for Claude Code and other MCP clients
- One-line pip install with automatic agent config updates
Behind the verdict
Coding agents forget, and that forgetfulness is expensive. Every new session re-explains your stack, your naming conventions, the architectural decision you argued about three weeks ago. TrueMemory targets exactly that pain and nothing broader: it's a memory engine that sits under Claude Code, Cursor, Codex, or Gemini CLI and lets each of them show up already knowing the codebase. We'd reach for it when privacy is non-negotiable. The whole design is a single file at ~/.truememory/memory.tm on your disk, no cloud calls, no per-query meter running. If your code can't leave the machine but you still want persisting context across sessions, this is the shape of tool you want. Pick it too if you want to verify the claims rather than take them. The arXiv paper, benchmark scripts, eval configs, and per-category results are public — 93.0% LoCoMo, 96.5% multi-hop — so a skeptical engineer can re-run the eval instead of reading a marketing page. That transparency is rarer than it should be in the memory-tool market. Where it bites: this is a terminal install with agent config edits and no GUI, so it's a developer tool, not something a PM will pick up. Memory is machine-local by design — no sync across devices, no shared memory between teammates. If two engineers need the same agent to remember the same project, TrueMemory isn't built for that. The closest alternative for that need is Mem0 or Supermemory, which sit in the hosted, cross-device, team-shared lane TrueMemory stays out of. The trade is reach for locality: call a hosted API and you get team sync and easier product integration; keep it local and you get privacy and no bill. Choose by which constraint is harder for you. One caveat on the numbers. The 55ms recall and 96.5% multi-hop claims are vendor-reported on
Models under the hood
as of 2026-10-25
Limitations
- Memory is stored in a single SQLite file on one machine, so it does not sync across devices — a second laptop starts empty.
- There is no GUI: interaction is via the CLI and your agent's config.
- The free terminal version is aimed at individual developers, and there's no hosted API for web or mobile apps in that build.
- The open-source version is described by the vendor as early-stage, with documentation that may be sparse and features that are still evolving.
- Product-embedding use cases require the closed-source, licensed-per-deployment version.
- The vendor's performance figures (93.0% LoCoMo, 96.5% multi-hop, ~55ms recall) come from its own published benchmark suite rather than independent third-party testing.
as of 2026-10-10
Researching TrueMemory? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is TrueMemory still active?
TrueMemory is live · free plan; paid plans on request · last verified 22 Oct 2026 · viability monitor.
Viability score
How well maintained and how widely used is TrueMemory? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: October 2026
How we score →What do users say about TrueMemory?
What people actually say about TrueMemory — is it worth it?
We scanned public community sources for TrueMemory on Sep 15, 2026 and could not establish that the discussion we found is about this tool rather than something else sharing its name. Only 2 of the posts we fetched could be positively tied to TrueMemory. Rather than publish a sentiment score built on the wrong subject, we publish nothing here and re-run the scan.
How we verified this
Every listed fact about TrueMemory— pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 22 Oct 2026. Pages mix model-assisted synthesis with human review; see how we verify.
Verification history
We have re-verified TrueMemory 10 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — changed SSL certificate
- — changed pricing tiers, summary, description, our verdict
- — changed switching in, switching out, who should skip it, summary, our verdict, features, hidden costs, setup time, description, pricing fit
- — changed integrations, features, who should skip it, who it suits, description, our verdict, summary, pricing tiers
- — changed link health
- — changed description, who it suits, our verdict, features, hidden costs, limitations, switching in, switching out, who should skip it, pricing tiers, pricing fit, setup time, summary, use cases
Showing the 6 most recent of 10 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does TrueMemory cost in 2026?
TrueMemory: Free to start; it also has paid plans whose price we could not confirm — see the pricing table. Pricing last verified 22 Oct 2026.
Plans compared
Every published TrueMemory tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| Free (terminal / coding version) | Free | — | — |
| Licensed (deployment) | Custom | quote | Product teams building ambient AI, legal-intelligence, or healthcare products that need the memory engine running inside their own deployment rather than called as an API.Adds: Closed-source, fully built deployment that ships inside your product instead of the CLI build, runs on your machine with no telemetry to a vendor cloud, and is scoped per engagement. |
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Where the pricing makes sense
The company stage and team size where TrueMemory's pricing actually pencils out — and where peers do it cheaper.
The coding version of TrueMemory is free for individual developers, which undercuts hosted memory layers that charge per query or per stored memory — Mem0 and Supermemory both sit in that hosted, cross-device territory. Teams that need shared memory across devices will find TrueMemory's free tier fits a solo dev's budget but not a team's requirements, and the licensed per-deployment option is scoped case by case for ambient, legal, and healthcare products.
Has TrueMemory's pricing changed?
No plan price for TrueMemory has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick TrueMemory?
Pick TrueMemory if you are Solo developers running Claude Code, Cursor, Codex, or Gemini CLI who want agent memory that persists between sessions, Privacy-conscious developers who need context to stay on the machine, with no cloud calls and no per-query bill, Researchers who want reproducible memory benchmarks with public scripts, eval configs, and per-category results, Developers on long-running projects where the same architectural decisions keep getting re-explained to fresh sessions.
Look elsewhere if you are Teams needing shared memory that syncs across devices or between teammates, Non-technical users — setup is a terminal install with agent config edits and no GUI, Web or mobile app developers who need a hosted memory API they can call from a product front end, Anyone who needs a memory layer on a machine they don't control, or telemetry into a vendor dashboard.
Skip it if Skip TrueMemory if you need memory that follows you across devices or between teammates, or if you want a hosted API you can call from a web or mobile front end instead of a CLI install.
GeneratedReal-world workflow fit
Concrete scenarios for the personas TrueMemory actually fits — and what changes day-one when you adopt it.
You pip install truememory, it detects Claude Code, Cursor, Codex, and Gemini CLI, attaches the MCP server and rewrites the agent configs. From there the surprise gate filters your sessions down to roughly 7.2% of observed input and writes it to ~/.truememory/memory.tm.
Outcome: Your next session opens with the agent already knowing your stack and prior architectural decisions, with recall reported at around 55ms and no cloud calls or per-query bill.
You run the pip install on a machine that never reaches the internet, and every memory stays in one local SQLite file with zero cloud calls.
Outcome: You get persistent agent recall without sending source code or context to a vendor, and a 30MB base footprint that won't crowd the disk.
You pull arXiv:2605.04897, clone the public benchmark scripts and eval configs, and re-run the LoCoMo evaluation against the per-category results the vendor published.
Outcome: You can verify or challenge the vendor's 93.0% LoCoMo and 96.5% multi-hop figures on your own hardware instead of trusting a marketing page.
Use cases
- Capture and recall context across multi-turn coding sessions in Claude Code without re-explaining your stack.
- Persist long-term memory for CLI agents across sessions with zero cloud costs or per-query bills.
- Support multi-hop and temporal reasoning on complex software projects where past decisions matter.
- Run fully offline memory for privacy-sensitive AI workflows on a local machine.
- Provide a memory layer for ambient AI that filters noise and retains only signal.
- Flag contradictions in long legal transcripts, as in the vendor's reported $1B+ case deployment.
- Maintain longitudinal patient context so today's symptoms can be checked against older diagnoses.
Setup time & first value
How long it actually takes to get something useful out of TrueMemory — broken out by persona, not the marketing-page minute.
Solo developer: about thirty seconds — a single pip install creates ~/.truememory/memory.tm, attaches the MCP server and updates configs for the CLI agents it detects. Researcher: minutes to install, longer to pull the benchmark scripts and re-run the LoCoMo eval configs. Product team shipping memory inside your own build: the licensed, closed-source deployment is scoped per engagement, with the
Switching to or from TrueMemory
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From manual context files or a CLAUDE.md-style notes file: install TrueMemory and let the surprise gate and salience layer capture what actually matters across sessions instead of you hand-maintaining a static document.
- →From a hosted memory layer like Mem0 or Supermemory: install TrueMemory locally if you want memory to live on your machine in ~/.truememory/memory.tm with no per-query bill, accepting that you give up cross-device and
- ↗To Mem0 or Supermemory: move if you need hosted, cross-device, team-shared memory rather than a single SQLite file on one machine.
Integrations
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “TrueMemory”, and we withheld 6: 6 could not be judged, because “TrueMemory” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about TrueMemory.
Official links
What are the best TrueMemory alternatives?
The closest alternatives to TrueMemory we list are Mem0, PMB and Jumbo.Cli. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| Mem0 | Freemium | Pricing not published | — |
| PMB | Free | Free | — |
| Jumbo.Cli | Freemium | Pricing not published | — |
Alternatives to TrueMemory
View allTools that pair well with TrueMemory
Common stack mates teams adopt alongside TrueMemory, with the specific reason each pairing earns its keep.
Mem0
Mem0 is a memory layer that gives AI agents and apps persistent context across sessions and users
PMB
Local-first memory layer that gives your AI coding agent a SQLite file it can recall from between sessions.
Jumbo.Cli
Jumbo.Cli gives AI coding agents persistent project memory from a local-first CLI, so you stop re-onboarding Claude Code every session.
TrueMemory vs its closest rival
TrueMemory's closest head-to-head is with Spider Cloud — TrueMemory and Spider Cloud solve complementary problems. If your pain point is AI forgetting context between sessions while using local coding agents, TrueMemory’s free, offline, SQLite-based memory is a no-brainer. If you need to inject fresh web data into your AI workflows, Spider Cloud’s high-speed, low-cost scraping API with new Browser AI commands is the clear choice. They are not direct competitors but can be used together. Verdict generated
Head-to-head comparisons
Truememory vs Spider Cloud
TrueMemory and Spider Cloud solve complementary problems. If your pain point is AI forgetting context between sessions while using local coding agents, TrueMemory’s free, offline, SQLite-based memory is a no-brainer. If you need to inject fresh web data into your AI workflows, Spider Cloud’s high-speed, low-cost scraping API with new Browser AI commands is the clear choice. They are not direct competitors but can be used together.
GeneratedTruememory vs Presto Voice
Choose Presto Voice if you operate a QSR chain and want to automate drive-thru ordering with proven revenue lift and upsell capabilities. Choose TrueMemory if you are a developer seeking a free, local, zero-cloud memory layer for AI coding agents to maintain persistent context across sessions. These tools solve entirely different problems and are not direct competitors.
GeneratedTruememory vs Temporal Ai
Choose Temporal AI if you need enterprise-grade durability, retries, and visibility for complex, long-running workflows or AI agents. Choose TrueMemory if you want a free, local, privacy-first memory layer for your CLI coding agents with zero cloud dependencies. For most AI agent memory needs, TrueMemory is simpler and cheaper; for orchestration of multi-step processes, Temporal is unmatched.
GeneratedFrequently asked questions
Topics
Used TrueMemory? Help shape our editorial sentiment research.