--- title: "Understand why Clawdbot forgets between sessions and keep what it loses in files" description: "Clawdbot, now OpenClaw, was built stateless for execution and security. See what the reset costs and where to keep the context it does not hold." canonical: "https://scalewithsearch.com/articles/clawdbot-forgets-context" date: "2026-01-28" modified: "2026-09-25" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # Understand why Clawdbot forgets between sessions and keep what it loses in files. Clawdbot moved your mouse. It typed in your terminal. It opened a browser tab, went to a page, and filled out a form, all from a voice note you sent through Telegram. The next day, you said, "Finish what we started yesterday." It had no idea what you meant. Peter Steinberger's open-source agent drew wide attention in January 2026. Public reports counted about 100,000 GitHub stars by early February 2026. It offered full computer control: mouse, keyboard, browser, shell commands, and messaging through WhatsApp, Telegram, Discord, iMessage, and Slack. The same reports stated that it did not carry anything from one session to the next. The project now calls itself OpenClaw, and its current repository lists sessions among its features. Test what your installed version keeps before you rely on any claim here. This article explains why the January 2026 design reset each session and what that reset costs. It also shows where to keep the context the agent does not hold. ## Know what Clawdbot does well Clawdbot solves a real problem. It takes AI out of the chat box and into the operating system. Most AI tools live in a browser tab. You type, they answer, and the output stays in that window. Clawdbot controls the computer the way you do. It opens applications, moves through interfaces, runs shell commands, and works with files. It closes the gap between an AI that talks and an AI that acts. The messaging channels make it easy to reach. You send a voice note through WhatsApp and get the result in Slack. The interface disappears. For single-session tasks, this is strong. It can automate a repetitive workflow, collect data from a website, or set up a development environment. It can also run a sequence of commands you would otherwise type. The problem appears when the work needs continuity. ## See where the session reset breaks the work You spend twenty minutes walking Clawdbot through your project structure, the client's requirements, the files involved, and your team's naming rules. It executes the task well. The next day, you start a new session. In the January 2026 version, the agent had no record of the project, the client, the rules, or the twenty minutes. You explain it all again. Or you skip the explanation, and the output comes back wrong because the agent works without context. The reset was not occasional. Every session started cold. For a quick automation, that is acceptable. For professional work where context builds over days, weeks, and months, it disqualifies the tool as the only layer. You cannot build a working relationship with a tool that starts over each time. ## Know why it was built without memory ### A stateless design The design was stateless on purpose. Each session was an isolated run: input goes in, actions happen, and the session ends. No storage layer connected one conversation to the next. No database logged your preferences, and no file recorded your history. That fits what Clawdbot was built to do: execute discrete tasks. It is a command-and-control agent, not a knowledge worker. The design chose breadth of action (control of the whole computer) over depth of context (knowledge of your business over time). Many open-source agents follow this pattern. A capable execution layer is hard enough to build without long-term memory as well. The two problems need different designs, and most teams choose one. ### A security trade-off Statelessness also has a real security argument, and it matters more for an agent with full system access. Clawdbot can control the mouse, keyboard, and terminal. Suppose it also kept a persistent store of your business details, client data, project structures, and workflow patterns. That store becomes a target. A compromised memory layer inside an agent with system-level permissions is a large attack surface. A stateless design avoids that. With no persistent data, nothing waits between sessions to be taken. Each session starts clean, and the security model has less to protect. That was a reasonable choice for an open-source tool aimed at wide adoption. It means an owner who needs persistent context must keep it somewhere else, and must decide how that store is protected. The [client data in agent memory article](/articles/client-data-in-ai-agent-memory) covers what belongs in such a store and what does not. ## Count what the reset costs The cost is not only the time spent on explanations. It is the difference in quality between work with context and work without it. An AI that knows your business produces different work from one that does not. With your client list, price structure, terms, voice, and past decisions, it produces output you can use. Without them, it produces the statistical average: generic, and right for nobody in particular. Measure the time with your own numbers. Suppose you spend 10 minutes per session re-establishing context and run three sessions a day. That is 30 minutes a day, or about two and a half hours a week. Multiply by your hourly rate to see the yearly total. The larger loss is compounding. When each session builds on the last, the AI becomes more useful over time. Without memory, that never starts, and each session stays at day one. The [article on re-explaining your business to AI](/articles/stop-re-explaining-your-business-to-ai) shows how to measure and remove that repetition. ## Keep the durable context in files The fix is not a smarter agent. It is a different place for the context. Claude Code does not try to make the model remember internally. It reads your local files. Your knowledge base lives on your machine in folders you control, and Claude Code reads the relevant files in each session. You do not paste anything by hand. The files stay on your disk, but the text Claude Code reads goes to Anthropic for processing. Retention terms depend on your plan. The anchor is one Markdown file, `CLAUDE.md`, at the root of the project folder. It holds your business context: who you are, your domains, your key clients and projects, your frameworks, your voice, and your output preferences. A well-built file runs 50 to 200 lines. You write it once and update it when the business changes. The difference shows in ordinary requests: | Request | Without the file | With the file | |---|---|---| | "Draft a follow-up to the Client A account." | A generic follow-up | A follow-up that knows the project and how you write to this client | | "Write this in my voice." | A default tone | Text that follows your voice rules | | "Use our standard proposal format." | An invented format | Your real proposal structure | There is no database, API, or infrastructure to run for this layer. The file sits in a folder, and the knowledge base grows as you add files beside it. ## Worked example: "finish what we started yesterday" Go back to the opening request. Write down what an agent would need to answer it: 1. What "we started yesterday": the task name and its goal. 2. Where the work stopped: the last completed step and the next step. 3. Which files and systems were involved. 4. Which rules applied: naming, client requirements, and limits. 5. What must not happen without approval, such as a send or a delete. Now put those five items in a file. At the end of each session, ask Claude Code to update `work/in-progress.md`. The update records the task, the last step, the next step, the files, and the open questions. Review the update. The next day, the request "finish what we started yesterday" has an answer on disk. Claude Code can do the work, or it can hand a clear instruction to Clawdbot. Test it once. Start a fresh session, give only that request, and check that the answer names the right task, the right next step, and the source file. If the file is missing, the correct result is a report that the file is missing, not a guess. ## Solve memory first, then add execution Clawdbot and Claude Code do not compete. Clawdbot gives the AI hands: it works with the operating system in ways that Claude Code does not try to. Claude Code with a context file gives the AI a record: it keeps your business context across sessions. The complete setup uses both, execution and context together. The [Clawdbot memory setup guide](/articles/clawdbot-ai-memory-setup) shows how to connect them with an approval step between draft and send. If you must solve one problem first, solve memory. Automation without context produces fast wrong output. Context without automation still produces accurate work that you can send by hand. The [Clawdbot and Claude Code comparison](/articles/clawdbot-vs-claude-code-computer-terminal-work) gives a bounded trial for choosing the execution surface once the context layer exists. ## Related: AI memory - [Give AI a context file so it knows your business without model training](/articles/ai-memory-vs-ai-training) - [Keep AI memory in local Markdown files instead of a cloud stack](/articles/ai-memory-without-cloud) - [Give Clawdbot the business context it loses between sessions](/articles/clawdbot-ai-memory-setup) ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```