# Memco

> Memco is shared memory for AI agents. It turns what an agent got wrong, what it got right and what changed into sourced, scoped lessons that later agents reuse — across models, tools and teams. Two domains: Coding, and Knowledge Work (research, client delivery, decisions and recurring operations).

Connect over MCP at https://spark.memco.ai/mcp. Product documentation lives at https://docs.memco.ai — that, not this site, is authoritative for commands, tool names and configuration.

Plans: Builder is free for one person; Teams is free for up to 4 people and paid from the fifth seat; Enterprise is licensed. See the pricing page for current terms.

## Start here

- [Memco — shared memory for AI agents](https://www.memco.ai/): Memco gives every agent access to what your company has already learned — shared memory that carries across models, tools and teams. 2.6× the tasks solved, same weights.
- [Product — how Memco works](https://www.memco.ai/product): Memco Shared Memory is a memory layer for AI agents: hybrid retrieval, Bayesian trust modelling and autonomous memory ops behind one MCP & CLI surface. Private by default, nothing pooled with anyone else, managed or self-hosted.
- [Knowledge Work — every task should make the next agent better](https://www.memco.ai/knowledge-work): Knowledge work is the research, analysis, client delivery, decisions and recurring operations where judgement matters. Memco turns corrections and decisions into sourced, scoped lessons that future agents reuse across models and tools.
- [Get Spark working in one real project](https://www.memco.ai/quickstart): Connect Spark over MCP to the agent you already use — Claude Code, Codex, Cursor or any MCP client — then search shared memory before you code and save what worked for the next agent.
- [Pricing](https://www.memco.ai/pricing): Builder is free for one person. Teams is free for up to 4 people, then $599 per user per year. Enterprise is custom for VPC, residency and governance. Memory operations stay unmetered for normal use.

## Use cases

- [Use cases](https://www.memco.ai/use-cases): Where shared agent memory lands first: engineering teams running coding agents on real repos, enterprises with boundaries and residency requirements, and partner programmes in review.
- [Memco for engineering teams](https://www.memco.ai/use-cases/engineering): Memco turns fixes, failed paths, repo quirks, review comments and human corrections into governed shared memory for every coding agent on your team. Same tools, same repos, same models — without the cold-start tax.
- [Memco for Customer Support | Shared Memory for Support Agents](https://www.memco.ai/use-cases/customer-support): Turn resolved tickets, escalation patterns, policy boundaries, and human corrections into governed memory your support agents can reuse.
- [Memco for Private Equity | Shared Memory for Portfolio AI Agents](https://www.memco.ai/use-cases/private-equity): Memco helps private equity firms turn agent deployments across portfolio companies into governed, reusable institutional memory.
- [Memco for Enterprise AI | Shared Memory for Enterprise Agents](https://www.memco.ai/use-cases/enterprise): Memco helps enterprises turn agent work into governed, reusable institutional memory across teams, models, tools, and workflows.
- [Enterprise — governed private memory](https://www.memco.ai/enterprise): Governed private memory for enterprise agent programmes: isolated scopes, provenance and audit on every lesson, deployment and residency options, and a pathfinder programme to start.

## Comparisons

- [Compare Memco with agent memory, context, search and skills platforms](https://www.memco.ai/compare): Compare Memco with Mem0, Zep, Cognee, Onyx, Sentra and Tessl by product job, learning lifecycle, proof question and enterprise control.
- [Memco vs Mem0: persistent agent memory vs governed organizational learning](https://www.memco.ai/compare/mem0): Compare Mem0 and Memco. See the difference between broad persistent memory infrastructure and evidence-backed organizational learning across people, agents and tools.
- [Memco vs Zep: temporal context vs engineering learning](https://www.memco.ai/compare/zep): Compare Zep and Memco. See the difference between temporal context infrastructure and evidence-backed lessons that improve later engineering work.
- [Memco vs Cognee: broad agent-memory platform vs packaged engineering learning](https://www.memco.ai/compare/cognee): Compare Cognee and Memco. See the difference between an open-source and managed agent-memory platform and a packaged engineering-learning workflow built around source-backed outcomes.
- [Memco vs Onyx: enterprise AI over connected knowledge vs governed learning](https://www.memco.ai/compare/onyx): Compare Onyx and Memco. See the difference between an enterprise AI interface with search, agents and actions over connected sources, and a governed learning loop built from real agent work.
- [Memco vs Sentra: company context graph vs governed engineering learning](https://www.memco.ai/compare/sentra): Compare Sentra and Memco. See the difference between a bi-temporal company brain with Code Memory and a governed engineering-learning loop built from corrections and outcomes.
- [Memco vs Tessl: governed skill artifacts vs governed memory lessons](https://www.memco.ai/compare/tessl): Compare Tessl and Memco. See the difference between turning learned practice into governed skills, plugins and CI checks, and carrying source-backed memory lessons from real work into later tasks.

## Research and writing

- [Research — memory, measured](https://www.memco.ai/research): The research behind shared agent memory: published papers with abstract-safe figures, the field guide, the journal archive and recorded talks. Every number carries its evidence class and source.
- [Agentic Engineering Memory — a Memco field guide](https://www.memco.ai/field-guide): A book-like reading surface on engineering memory for agent programmes: preface, plates, fifteen chapters and a diagnostic — what a year of building taught us about making coding agents compound, not repeat.
- [Journal — memory, measured](https://www.memco.ai/blog): Research, engineering practice and field notes on how agents learn from real work — and how that learning becomes reusable company memory. All 30 essays.
- [Videos](https://www.memco.ai/blog/videos): Memco Journal videos: product walkthroughs, market commentary, build sessions, and customer conversations.

## Company

- [About](https://www.memco.ai/about): Memco builds the memory layer agents run on. The thesis, three convictions, the founders, and open roles in London, San Francisco and Stockholm.
- [Careers](https://www.memco.ai/careers): Join Memco to build the memory layer for AI agents. Three open roles across engineering, design and go-to-market — remote, load-bearing from day one. Open applications welcome.
- [Partner with Memco — channel and referral partnerships](https://www.memco.ai/partners): Introduce Memco to the teams you advise and earn on every introduction that becomes a customer. Commission on first-year revenue, deal registration, no quotas and no resale obligations.
- [Brand — the wordmark, the palette, the type](https://www.memco.ai/brand): Everything you need to represent Memco: the wordmark and monogram as vectors, the colour palette, the three typefaces, and the rules that come with them.

## More

- [agents.md — setup instructions for AI agents](https://www.memco.ai/agents): Written for AI agents, not their humans: what Memco changes about how you work, and the exact steps to get your human set up with an account, the CLI and MCP.

## Optional

- [Learning on the Job, Batteries Included](https://www.memco.ai/blog/learning-on-the-job-batteries-included): An agent learns a business's unwritten rules and compliance climbs from 20% to 64%. The harness behind our Learning on the Job paper is now open source, so you can reproduce that curve for about $3.
- [Memory for Knowledge Work](https://www.memco.ai/blog/memory-for-knowledge-work): Your teams correct AI assistants every day, but that knowledge disappears with the session. Spark's new Knowledge Work domain turns those corrections into governed organisational memory every assistant can reuse.
- [Sovereign AI Needs Sovereign Memory](https://www.memco.ai/blog/sovereign-ai-needs-sovereign-memory): Sovereign AI is more than controlling where the model runs. New research shows why organisations should also own the governed, portable memory that gives agents their operational knowledge.
- [Learning on the Job: Agents Can Learn from the Feedback They Already Get](https://www.memco.ai/blog/learning-on-the-job): Every business has feedback its AI agents can learn from. New research shows that external memory turns routine corrections into portable, compounding operational knowledge.
- [Coding agents are the wedge. Organizational memory is the prize.](https://www.memco.ai/blog/coding-agents-are-the-wedge-organizational-memory-is-the-prize): Coding agents are the cleanest place to see the agent-memory problem: every run creates evidence, but without trusted organizational memory the next agent starts cold.
- [Knowledge Management Systems Are Always Obsolete. Agents Can Fix That](https://www.memco.ai/blog/kms-for-agents): Every enterprise knows its knowledge base is out of date. AI agents can finally fix that, but only if the memory infrastructure meets enterprise requirements. We walk through what it takes: identity and access control, knowledge scoping, human oversight, provenance, and data residency.
- [Why Active Agentic Memory is the Next Shift](https://www.memco.ai/blog/the-zig-and-zag-of-ai): Human-crafted knowledge works perfectly fine, until data-driven learning surpasses it. Explore the historical "Zig and Zag" of AI, and why I believe shared agentic memory is the infrastructure required for the next era of autonomous learning.
- [Your Team Knows More Than Anyone On It](https://www.memco.ai/blog/your-team-knows-more-than-anyone-on-it): Most AI memory tools give back what you put in. With Knowledge Abstraction, Spark derives principles your team never stated, and helps agents avoid problems nobody has encountered yet.
- [Your Agent's Memory Is a Markdown File. That's a Problem.](https://www.memco.ai/blog/your-agent-memory-is-a-markdown-file): Files are the most popular form of agent memory. Here are four things they structurally cannot do.
- [Your Coding Agent Remembers Everything, Until It Doesn't](https://www.memco.ai/blog/your-coding-agent-remembers-everything-until-it-doesnt): Context windows can retrieve facts with remarkable accuracy, until compaction silently erases the project rules and conventions your team depends on.
- [Continual Learning for Enterprise AI Needs a Memory Layer](https://www.memco.ai/blog/continual-learning-for-enterprise-ai-needs-a-memory-layer): Most enterprise AI systems do not fail because the model is incapable. They fail because the system cannot retain, refine, and reuse what the organization has already learned.
- [Making Sense of Agentic Memory: A Map of the Design Space](https://www.memco.ai/blog/making-sense-of-agentic-memory): The term 'memory' is used to describe wildly different systems. We map out the design space and explain the choices behind Spark's shared memory architecture.
- [If Your Coding Agents Don't Share Memory, You're Burning Money](https://www.memco.ai/blog/if-your-coding-agents-dont-share-memory-youre-burning-money): Most teams running AI coding agents pay for the same knowledge over and over. We ran 200+ evaluation runs to show how shared memory cuts token usage up to 87% and lifts pass rates from 70% to 100%.
- [What Happens When Your AI Agent Fails?](https://www.memco.ai/blog/what-happens-when-your-ai-agent-fails): What happens when an agent fails? We analyzed the ROI of Spark shared memory and discovered that costs decrease by 40% on average. The pleasant surprise was the 'fail cheap': even unsolved tasks see a cost reduction of 34%.
- [Agents need knowledge they can't generate themselves](https://www.memco.ai/blog/agents-need-knowledge-they-cant-generate): A new paper confirms what we've been building toward: curated procedural knowledge improves agent performance by 16%, but agents can't write it themselves. Static skill files don't scale. Shared memory does.
- [AI-Generated Code Might Be More Maintainable Than You Think](https://www.memco.ai/blog/ai-generated-code-might-be-more-maintainable-than-you-think): A preprint on AI and software maintainability finds no significant difference in downstream maintainability—and habitual AI users may produce slightly cleaner code. Caveats and takeaways.
- [Why Shared Memory Matters Even for "Solved" Problems](https://www.memco.ai/blog/why-shared-memory-matters-even-solved-problems): When an agent already solves a task 100% of the time, does memory help? Our experiment with bug-fixing in Go shows Spark cuts cost and variance in half—even on the easy problems.
- [Why AI Struggles with Your Legacy Code](https://www.memco.ai/blog/why-ai-struggles-your-legacy-code): The gap between AI demos and real engineering teams: legacy code, internal APIs, and tech debt. How a shared agentic memory closes that gap with collective continual learning.
- [Beyond the Human Ceiling](https://www.memco.ai/blog/beyond-human-ceiling): AI is surpassing human performance in specialized domains. In software development, agentic knowledge-sharing infrastructure like Spark can outperform the human forums it replaces.
- [Smarter Together: Our New Paper Shows How Shared Memory Lifts All Models](https://www.memco.ai/blog/smarter-together-shared-memory-lifts-all-models): Our first evaluation paper demonstrates that access to Spark's shared, active memory benefits AI coding agents of all sizes—from smaller open-weights models to the largest commercial models.
- [The Collective Intelligence Gap](https://www.memco.ai/blog/collective-intelligence-gap): We're building an agentic workforce. Now we need to build the infrastructure for collective intelligence.
- [The Wedge and the Vision: Code to Agentic Infrastructure](https://www.memco.ai/blog/wedge-vision-code-agentic-infrastructure): LLMs broke software's most powerful feedback loop — learning from distributed experience. Shared memory can fix it. Our path from code agents to the runtime for agentic systems.
- [79% of Enterprises Are Deploying AI Agents. Only One Thing Is Missing.](https://www.memco.ai/blog/79-percent-enterprises-deploying-ai-agents-only-thing-missing): Enterprises succeeding with AI aren't the ones with the most powerful models. They're the ones that figured out how to make organizational knowledge compound through private shared memory networks.
- [What Developers Actually Want From AI Memory: Lessons from 230+ HN Comments](https://www.memco.ai/blog/what-developers-actually-want-from-ai-memory): The Hacker News response to Claude's Memory launch reveals a sharp divide: sophisticated users want sovereignty, control, and learning - not accumulation. Here's what the community is demanding.
- [Karpathy is Talking About Cognitive Core, We're Building the Memory for It](https://www.memco.ai/blog/karpathy-cognitive-core-memory): Andrej Karpathy's vision of separating intelligence from knowledge aligns with our R&D at Memco. Our Spark memory layer enables smaller models to achieve state-of-the-art performance by externalizing specialized knowledge.
- [Memory Incentives and Architectures](https://www.memco.ai/blog/memory-incentives-and-architectures): As AI performance gains slow, memory becomes the new battleground. We explore service-provider, portable, and community memory architectures, and how they incentivize learning and ownership.
- [Nature's Blueprint: What Human Memory Teaches Us About Building Smarter AI Agents](https://www.memco.ai/blog/natures-blueprint-human-memory-ai-agents): By studying biological memory, we can design more effective memory systems for AI agents that learn, generalize, and strategically forget
- [Reinforcement Learning For AI Agents: Learning on the Job with Active Memory](https://www.memco.ai/blog/active-memory-reinforcement-learning-for-agents): How an active memory system enables runtime reinforcement learning, transforming AI agents from static tools into dynamic partners that learn from experience
- [The Art of Forgetting: Why True Memory is the Next Frontier for Autonomous AI Agents](https://www.memco.ai/blog/the-art-of-forgetting): Building truly autonomous agents requires moving beyond simple retrieval to design active memory systems that can learn, generalise, and strategically forget
- [Introducing Memco: the shared memory layer for AI agents](https://www.memco.ai/blog/welcome-to-memco): Every day, thousands of developers and their AI agents solve the same problems—then forget the solutions. We're building Memco to fix that.
- [Agents That Learn](https://www.memco.ai/blog/agents-that-learn): Why collective learning and shared memory are the missing pieces for AI agents
- [Your Next User Is Not Human](https://www.memco.ai/blog/your-next-user-is-not-human): The Developer Flywheel is Broken. How AI agents are replacing humans as the primary consumers of developer tools and documentation.
- [Cross-agent continual learning from minimal feedback](https://www.memco.ai/blog/cross-agent-continual-learning)