Chat, Vibe and get work done without internet, all in one session. Assistant to get things done with your files and projects.
✓
Stays private
Prompts never leave your device. No account, no cloud, nothing to leak.
✓
Stays light
Small on-device model tuned for everyday CPUs. No GPU required.
✓
Works offline
Internet needed once for setup — after that, fully unplugged.
Questions, answered
The short version. Everything else is in the app.
Get Arch for your machine
Two installers: Windows (.exe with guided setup) and macOS (wizard that builds Arch.app). No account needed. Pick your system and you're chatting in minutes.
These Terms and Conditions ("Terms") govern your access to, download of, and use of Arch (the "Software"), an open-source, self-hosted coding agent made available at github.com/Wizzit-254/Arch-Assistant-Repo (the "Repository"). These Terms are entered into between you ("you," "your," or "User") and the maintainer(s) of Arch ("we," "us," "our," or the "Maintainer").
By downloading, installing, running, forking, modifying, or otherwise using the Software, you agree to be bound by these Terms. If you do not agree to these Terms, you must not download, install, or use the Software.
If you are using the Software on behalf of an organization, you represent that you have the authority to bind that organization to these Terms, and "you" will refer to that organization as well as yourself individually.
1.1 Relationship to the software license
These Terms govern your conduct and legal relationship with us as a user of the project and its associated infrastructure (Repository, issue tracker, communications). They are separate from, and do not replace, the open-source license under which the Software's source code is distributed (see Section 3). Where these Terms and the license conflict regarding your rights to use, copy, modify, or distribute the source code itself, the license controls.
2. Definitions
3. Open-Source License Grant
The Software's source code is distributed under the MIT License, a copy of which is included in the Repository (typically as a LICENSE file). The MIT License governs your rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, subject to the conditions stated in that license (including retention of the copyright notice and license text).
Nothing in these Terms is intended to restrict any right granted to you under the MIT License. In the event of any direct conflict between these Terms and the MIT License regarding your rights to the source code itself, the MIT License controls. These Terms instead address matters the license does not cover, such as acceptable use of the Repository, disclaimers specific to AI-generated Output, and your relationship with the Maintainer as a user or contributor.
3.1 No additional restrictions on the licensed code
We will not impose additional restrictions on your exercise of the rights granted under the MIT License. However, if you use the Software in a hosted, commercial, or redistributed form, you remain solely responsible for your own compliance with the MIT License's conditions (such as including the original copyright and license notice).
3.2 Trademarks
This license grant does not include any right to use the Arch name, logo, or branding in a way that suggests endorsement or affiliation with the Maintainer without prior written permission, except as reasonably necessary to describe the origin of the Software (e.g., "based on Arch").
4. Eligibility and Acceptable Use
4.1 Eligibility
You must have the legal capacity to enter into a binding agreement in your jurisdiction to use the Software. The Software is intended for use by developers and technical users; it is not directed at children, and we do not knowingly permit use by individuals below the age of legal majority without appropriate parental or guardian consent where required by applicable law.
4.2 Acceptable use
You agree not to use the Software to:
Develop, generate, or facilitate malware, ransomware, exploits, or other tools designed to gain unauthorized access to systems or data.
Violate any applicable law, regulation, or third party's intellectual property, privacy, or other legal rights.
Generate or propagate content that is unlawful, fraudulent, defamatory, or that facilitates harassment, discrimination, or violence.
Misrepresent Output as being free of errors, security vulnerabilities, or legal issues where you have not independently verified it.
Circumvent, disable, or interfere with any security-related features of the Software or of any Third-Party Component it integrates with.
Use the Repository's issue tracker, discussions, or communication channels to spam, harass, or distribute unrelated commercial content.
4.3 Consequences of misuse
Because the Software is self-hosted and we do not operate the infrastructure on which it runs, we generally have no visibility into, and no ability to prevent, your use of the Software once downloaded. Your compliance with this Section is a matter of your agreement with us and, where relevant, applicable law — not something we can enforce technically. We may, however, moderate the Repository itself (see Section 5.3) and may pursue any remedy available to us at law for violations of these Terms.
5. Self-Hosted Software; No Hosted Service
Arch is distributed exclusively as self-hosted, open-source software. We do not operate, and these Terms do not create, any hosted, managed, or software-as-a-service ("SaaS") version of the Software. There is no account system, subscription, uptime commitment, or service level agreement associated with the Software itself.
5.1 Your infrastructure, your responsibility
You are solely responsible for:
Provisioning and securing the hardware, operating system, and network on which you run the Software.
Configuring the Software appropriately for your environment, including any Third-Party Components you choose to connect.
Maintaining backups of your code, data, and configuration independent of the Software.
Any costs associated with running the Software, including compute, storage, and any third-party API usage fees you incur.
5.2 No service relationship
Because we do not host or operate the Software on your behalf, no service-provider relationship, agency relationship, or fiduciary duty is created between you and the Maintainer by virtue of your use of the Software.
5.3 Repository and community infrastructure
While the Software itself is self-hosted, the Repository (issue tracker, discussions, releases) is infrastructure we do administer, typically via a third-party platform (e.g. GitHub). We reserve the right to moderate, edit, or remove content from the Repository, and to restrict access for any user who violates these Terms, the platform's own terms of service, or applicable law.
6. Third-Party Models, Dependencies, and Integrations
6.1 Model-agnostic design
Arch can be configured to run with locally-hosted models or, at your option, with third-party large language model ("LLM") APIs and other Third-Party Components (such as version control hosts or MCP-compatible servers). The Maintainer does not operate, control, or warrant the availability, accuracy, or behavior of any Third-Party Component.
6.2 Your responsibility for third-party terms
If you configure the Software to connect to a Third-Party Component, you are solely responsible for:
Reviewing and complying with that third party's terms of service, acceptable use policy, and privacy policy.
Any fees, quotas, or rate limits imposed by that third party.
Ensuring you have the necessary rights or licenses to submit your code, prompts, or data to that third party.
Any consequences arising from that third party's handling, retention, or use of data you transmit to it, including for model training, unless that party's terms state otherwise.
6.3 Open-source dependencies
The Software incorporates open-source libraries and packages maintained by third parties, each governed by its own license. A list of dependencies and their licenses is typically available in the Repository's package manifest or a NOTICE/THIRD-PARTY-LICENSES file. You are responsible for complying with those licenses if you redistribute the Software or derivative works.
6.4 No endorsement
Compatibility with, or documentation referencing, a particular Third-Party Component does not constitute an endorsement of that component, its accuracy, its safety, or its business practices.
6.5 Interruption or discontinuation of third-party services
Third-Party Components you connect to the Software may change, restrict, or discontinue their services at any time, for reasons outside our control. We are not responsible for any disruption to your use of the Software that results from a change, outage, deprecation, or policy change by a Third-Party Component.
6.6 API keys and credentials
Where the Software requires you to supply credentials (such as API keys) to use a Third-Party Component, you are solely responsible for safeguarding those credentials. The Software is designed to store such credentials locally per your configuration; you should treat them with the same care as any other secret and avoid committing them to version control or otherwise exposing them.
7. User-Generated Content and Output
7.1 Your input
You retain all rights, title, and interest in any code, prompts, files, or other content you provide to the Software ("Input"). Because the Software is self-hosted with no telemetry by default, your Input is not transmitted to the Maintainer through ordinary operation of the Software (see our separate Privacy Policy for details).
7.2 Output generated by the Software
The Software may generate code, suggestions, or other Output based on your Input, using either a locally-run model or a Third-Party Component you configure. You are solely responsible for reviewing, testing, and validating any Output before relying on it, incorporating it into a production system, or distributing it.
7.3 No guarantee of originality or non-infringement
Output is generated algorithmically and may, in some cases, closely resemble pre-existing code, including code subject to third-party copyright, patent, or license restrictions. We do not warrant that Output is original, accurate, secure, fit for any particular purpose, or free from third-party intellectual property claims. You are responsible for independently assessing whether Output is safe to use in your context, including its licensing implications.
7.4 Ownership of output
As between you and the Maintainer, we make no claim of ownership over Output generated for you through your use of the Software. Any rights you hold in Output are subject to the terms of any Third-Party Component you used to generate it (some LLM providers impose their own terms regarding ownership or usage rights over model output).
8. Intellectual Property
8.1 Ownership of the Software
Subject to the rights granted under the MIT License, the Maintainer and other contributors retain their respective copyright and other intellectual property rights in the Software. The MIT License grants you specific, broad rights to use, modify, and redistribute the Software — it does not transfer ownership.
8.2 Trademarks and branding
Arch's name, logo, and any associated branding (to the extent used) are owned by the Maintainer and are not licensed under the MIT License. You may use them only as reasonably necessary to identify the Software's origin, or with our prior written permission for other uses.
8.3 Feedback
If you provide feedback, suggestions, or ideas about the Software outside of a formal Contribution (Section 9), you grant us a perpetual, irrevocable, royalty-free license to use that feedback for any purpose, without obligation to you, unless you expressly agree otherwise in writing.
8.4 Third-party intellectual property
Nothing in these Terms grants you rights in any Third-Party Component's intellectual property beyond what that component's own license or terms provide.
9. Contributions to the Project
9.1 Submitting Contributions
If you submit a Contribution (e.g. via a pull request, patch, or issue containing code) to the Repository, you agree that:
Your Contribution is licensed to the project under the same MIT License that governs the Software, unless the Repository's contribution guidelines specify otherwise.
You have the right to submit the Contribution under that license — that is, you either wrote it yourself, or you have the necessary rights and permissions to contribute it.
Your Contribution does not knowingly infringe any third party's intellectual property rights.
The Maintainer may use, modify, distribute, and relicense your Contribution as part of the Software, consistent with the governing open-source license.
9.2 No obligation to accept Contributions
We are under no obligation to review, accept, merge, or maintain any Contribution. Acceptance of a Contribution is at the sole discretion of the Maintainer or designated project reviewers.
9.3 Code of conduct
Contributors and community members are expected to engage respectfully with one another. If the Repository maintains a separate Code of Conduct document, it is incorporated into these Terms by reference for purposes of Repository participation.
9.4 Attribution
We will make reasonable efforts to preserve contributor attribution (e.g., via git commit history) but do not guarantee that every Contribution will be individually credited outside of standard version control records.
10. No Warranty
THE SOFTWARE IS PROVIDED "AS IS" AND "AS AVAILABLE," WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, TITLE, AND NON-INFRINGEMENT. TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THE MAINTAINER DISCLAIMS ALL WARRANTIES, WHETHER EXPRESS, IMPLIED, STATUTORY, OR OTHERWISE, INCLUDING ANY WARRANTIES ARISING OUT OF A COURSE OF DEALING OR USAGE OF TRADE.
Without limiting the foregoing, we do not warrant that:
The Software will be error-free, uninterrupted, or free of security vulnerabilities.
Any Output generated by the Software will be accurate, complete, secure, non-infringing, or suitable for your intended use.
Defects will be corrected, or that the Software will be compatible with any particular hardware, operating system, or Third-Party Component.
The Software or its use will meet any regulatory, compliance, or certification requirement applicable to your industry or jurisdiction.
This disclaimer is consistent with the "as is" basis on which open-source software is customarily made available and, where applicable, mirrors the warranty disclaimer already contained in the MIT License.
11. Limitation of Liability
TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, IN NO EVENT WILL THE MAINTAINER OR ANY CONTRIBUTOR BE LIABLE FOR ANY INDIRECT, INCIDENTAL, SPECIAL, CONSEQUENTIAL, EXEMPLARY, OR PUNITIVE DAMAGES, OR FOR ANY LOSS OF PROFITS, REVENUE, DATA, GOODWILL, OR BUSINESS OPPORTUNITY, ARISING OUT OF OR RELATED TO YOUR USE OF, OR INABILITY TO USE, THE SOFTWARE, WHETHER BASED ON WARRANTY, CONTRACT, TORT (INCLUDING NEGLIGENCE), OR ANY OTHER LEGAL THEORY, EVEN IF THE MAINTAINER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES.
TO THE MAXIMUM EXTENT PERMITTED BY APPLICABLE LAW, THE MAINTAINER'S TOTAL AGGREGATE LIABILITY ARISING OUT OF OR RELATED TO THESE TERMS OR THE SOFTWARE SHALL NOT EXCEED THE GREATER OF (A) THE AMOUNT YOU PAID, IF ANY, TO THE MAINTAINER FOR THE SOFTWARE IN THE TWELVE (12) MONTHS PRECEDING THE CLAIM, OR (B) ONE HUNDRED UNITED STATES DOLLARS (USD $100), GIVEN THAT THE SOFTWARE IS PROVIDED FREE OF CHARGE UNDER AN OPEN-SOURCE LICENSE.
11.1 Basis of the bargain
You acknowledge that the Maintainer has made the Software available to you free of charge, on an open-source, self-hosted basis, and that the limitations of liability in this Section are a fundamental basis of the bargain between you and the Maintainer, without which the Software would not be made available to you on these terms.
11.2 Jurisdictional limits
Some jurisdictions do not allow the exclusion or limitation of certain damages or implied warranties. In such jurisdictions, the exclusions and limitations in Sections 10 and 11 apply only to the maximum extent permitted by applicable law, and our liability will be limited to the smallest extent permitted.
12. Indemnification
To the maximum extent permitted by applicable law, you agree to defend, indemnify, and hold harmless the Maintainer and any other contributors from and against any claims, liabilities, damages, losses, and expenses, including reasonable legal fees, arising out of or in any way connected with:
Your use or misuse of the Software, including any Output you generate, use, or distribute.
Your violation of these Terms or of any applicable law or regulation.
Your violation of any third party's rights, including intellectual property, privacy, or contractual rights, in connection with your use of the Software or any Third-Party Component.
Any Contribution you submit that infringes a third party's rights or breaches the representations in Section 9.1.
We reserve the right, at our own expense, to assume the exclusive defense and control of any matter otherwise subject to indemnification by you, in which case you agree to cooperate with our defense of such claim.
13. AI-Specific Disclosures and Limitations
13.1 Nature of AI-generated Output
Arch uses large language models to generate code, suggestions, and other Output. AI-generated Output is inherently probabilistic and may be inaccurate, incomplete, insecure, or inconsistent, even when it appears confident or well-formed. You should treat all Output as a draft or suggestion requiring your own review, not as verified, production-ready work.
13.2 No professional advice
Output generated by the Software does not constitute professional advice of any kind (including legal, security, or engineering advice), even where it addresses topics related to those fields. Use of Output in contexts requiring professional judgment (e.g., security-critical, safety-critical, or regulated systems) should be independently reviewed by a qualified professional.
13.3 Known limitations of LLM-based tools
Depending on the model you configure, Output may exhibit limitations common to LLM-based coding tools, including but not limited to:
Hallucinated APIs, functions, or package names that do not exist.
Introduction of security vulnerabilities, including insecure defaults or outdated practices.
Incomplete handling of edge cases, error conditions, or concurrency issues.
Verification loops that pass local tests without guaranteeing correctness in your broader system.
Inconsistent behavior across sessions, model versions, or Third-Party Component updates.
13.4 Verification loops are not a warranty
Where the Software includes automated verification steps (such as running tests or checks before presenting Output), these mechanisms are designed to improve reliability but do not constitute a warranty or guarantee of correctness, security, or fitness for purpose. You remain responsible for your own testing and code review practices.
13.5 Training data and model behavior
If you connect a Third-Party Component (e.g. a hosted LLM), that model's training data, biases, and behavior are determined by the third-party provider, not by the Maintainer. We make no representation about the accuracy, currency, or appropriateness of any third-party model's training data or outputs.
13.6 No liability for reliance on Output
To the maximum extent permitted by applicable law, the Maintainer is not liable for any damages, losses, or claims arising from your reliance on Output generated by the Software, including but not limited to security incidents, data loss, financial loss, or intellectual property claims resulting from code suggested, generated, or modified by the Software.
13.7 Human oversight recommended
We strongly recommend maintaining human review over any code, configuration, or command that the Software proposes to execute or apply, particularly for actions with irreversible effects (such as deleting files, modifying production systems, or executing shell commands).
14. Security and Responsible Disclosure
14.1 Your security responsibilities
Because the Software is self-hosted, the security of your deployment (including any credentials, API keys, and code it processes) is your responsibility. We recommend following standard secure development practices, including keeping the Software and its dependencies up to date, and reviewing the Software's permissions before granting it access to sensitive systems or credentials.
14.2 Reporting vulnerabilities
If you discover a security vulnerability in the Software, please report it responsibly by emailing trevorkisingu@gmail.com rather than disclosing it publicly, to give us a reasonable opportunity to address it before public disclosure. We will make reasonable efforts to acknowledge and investigate credible reports.
14.3 No guarantee of a security response timeline
As an open-source project maintained on a best-effort basis, we do not guarantee any specific response time, patch timeline, or resolution for reported vulnerabilities, though we will make reasonable efforts consistent with the resources available to the project.
15. Export Control and Sanctions Compliance
The Software may be subject to export control laws and regulations, including those of the jurisdictions in which it is developed, hosted, or accessed. By using the Software, you represent that you are not located in, under the control of, or a national or resident of any country or region subject to comprehensive trade sanctions, and that you are not listed on any applicable restricted-party or denied-persons list.
You agree not to use, export, re-export, or transfer the Software in violation of any applicable export control or sanctions law. You are solely responsible for ensuring your use of the Software, and any Third-Party Component you connect to it, complies with export control laws applicable to you.
16. Compliance with Applicable Law
You are solely responsible for ensuring that your use of the Software, including any Output it generates, complies with all laws and regulations applicable to you, including but not limited to intellectual property law, data protection law, employment law (where Output is used in automated decision-making), and any industry-specific regulatory requirements.
Where your use of the Software involves processing personal data (for example, if you configure it to analyze code containing personal data, or connect it to systems that do), you are responsible for ensuring that processing complies with applicable data protection laws, such as Kenya's Data Protection Act, 2019, the EU General Data Protection Regulation, or other applicable regimes, as relevant to your circumstances.
17. Termination
17.1 Termination by you
You may stop using the Software, and delete any local copies, at any time. Because there is no account or subscription associated with the Software, no further action is required to "terminate" your use of it.
17.2 Termination or restriction by us
We may, at our discretion, restrict your access to the Repository or related community infrastructure (e.g. by blocking an account from the issue tracker) if you violate these Terms, without affecting your rights to use previously downloaded copies of the Software under the MIT License.
17.3 Effect of termination
Termination of your access to the Repository or community infrastructure does not revoke the rights already granted to you under the MIT License with respect to code you have already obtained. Sections of these Terms that by their nature should survive termination will survive, including Sections 10, 11, 12, and 13 (see also Section 28, Survival).
18. Modifications to These Terms
We may revise these Terms from time to time to reflect changes in the project, applicable law, or our practices. When we make material changes, we will update the "Effective date" at the top of this document and, where practical, note the change in the Repository's changelog or release notes.
Continued use of the Software after an update to these Terms constitutes your acceptance of the revised Terms. If you do not agree to a revision, your remedy is to stop using the Software; your rights to previously obtained copies of the Software under the MIT License are unaffected by a change to these Terms.
19. Governing Law and Jurisdiction
These Terms, and any dispute arising out of or relating to them or to the Software (to the extent not governed by the MIT License itself), are governed by the laws of the Republic of Kenya, without regard to its conflict-of-laws principles.
Subject to Section 20 (Dispute Resolution), you agree that any legal action or proceeding arising out of or relating to these Terms shall be brought exclusively in the courts located in Kenya, and you consent to the personal jurisdiction of such courts.
20. Dispute Resolution
20.1 Informal resolution first
Before initiating any formal legal proceeding, you agree to first contact us at trevorkisingu@gmail.com and attempt in good faith to resolve the dispute informally. We will attempt to do the same. Most concerns can be resolved this way, particularly given the project's open-source, community-driven nature.
20.2 Arbitration (optional mechanism)
If a dispute cannot be resolved informally within sixty (60) days, either party may elect to refer the dispute to arbitration seated in Kenya, conducted in English, under the Arbitration Act (Cap. 49 of the Laws of Kenya) or such rules as the parties may mutually agree, in lieu of court proceedings, except where either party seeks injunctive or equitable relief to protect intellectual property rights, in which case that party may proceed directly to the courts described in Section 19.
20.3 Class action waiver
To the extent permitted by applicable law, disputes will be resolved on an individual basis, and you agree not to bring or participate in any claim as a plaintiff or class member in any purported class, collective, or representative proceeding against the Maintainer.
21. Severability and Waiver
If any provision of these Terms is found by a court or tribunal of competent jurisdiction to be invalid, illegal, or unenforceable, that provision will be limited or eliminated to the minimum extent necessary so that the remaining provisions of these Terms will remain in full force and effect.
No waiver of any term of these Terms will be deemed a further or continuing waiver of such term or any other term, and our failure to assert any right or provision under these Terms will not constitute a waiver of such right or provision.
22. Assignment
You may not assign or transfer these Terms, by operation of law or otherwise, without our prior written consent. We may assign or transfer these Terms, in whole or in part, without restriction, including in connection with a transfer of stewardship of the project to a new maintainer or organization. These Terms will bind and inure to the benefit of the parties, their successors, and permitted assigns.
23. Entire Agreement
These Terms, together with the MIT License, our Privacy Policy, and any Repository-specific guidelines referenced herein (such as a Code of Conduct or Contributing Guide), constitute the entire agreement between you and the Maintainer regarding your use of the Software, and supersede any prior or contemporaneous agreements, communications, or understandings, whether written or oral, regarding that subject matter.
24. Relationship to the Open-Source License
In the event of any conflict between these Terms and the MIT License specifically with respect to your rights to use, copy, modify, merge, publish, distribute, sublicense, or sell the source code, the MIT License governs. These Terms govern all other aspects of your relationship with the Maintainer, including disclaimers, liability limitations, acceptable use of the Repository, and AI-specific disclosures, which are not addressed by the MIT License.
25. Relationship to the Privacy Policy
Our separate Privacy Policy describes how data is (and, in most respects, is not) collected or processed in connection with the Software, consistent with its self-hosted, no-telemetry-by-default design. The Privacy Policy is incorporated into these Terms by reference. In the event of a direct conflict between these Terms and the Privacy Policy regarding the handling of personal data, the Privacy Policy governs on that specific topic.
26. Force Majeure
We will not be liable for any failure or delay in the maintenance, availability, or support of the Software resulting from causes beyond our reasonable control, including but not limited to acts of God, natural disaster, war, terrorism, riots, civil unrest, government action, labor disputes, internet or infrastructure failures, or failures of Third-Party Components.
27. Notices
We may provide notices to you regarding these Terms via the Repository (e.g., release notes, a pinned issue, or the README), or, where you have contacted us directly, via email. You may send notices to us at trevorkisingu@gmail.com. Notices are deemed given when posted to the Repository or, for email, when sent, absent evidence of non-delivery.
28. Survival
Provisions of these Terms that by their nature should survive termination of your use of the Software will survive, including without limitation Sections 3 (License Grant, to the extent already exercised), 8 (Intellectual Property), 9.1 (Contributions), 10 (No Warranty), 11 (Limitation of Liability), 12 (Indemnification), 13 (AI-Specific Disclosures), 19 (Governing Law), 20 (Dispute Resolution), and this Section 28.
29. Interpretation
Section headings in these Terms are for convenience only and do not affect interpretation. The words "including" and "such as" mean "including without limitation." These Terms have been made available in English; any translation is provided for convenience only, and the English version governs in the event of any inconsistency.
30. Contact Information
If you have questions about these Terms, wish to report a security vulnerability, or need to contact us for any other reason related to Arch, please reach out through:
We aim to respond to inquiries within a reasonable timeframe, consistent with the resources available to an open-source project maintained on a volunteer or part-time basis.
Arch is an open-source, locally-run coding agent. This policy explains what data Arch does and does not handle, how the software is designed with respect to your privacy, and what applies to you differently if you interact with the project's community infrastructure (such as its repository host, issue tracker, or any optional hosted services) rather than the software itself.
Because Arch is open-source and, in its default configuration, runs entirely on your own machine, this policy is shorter and more specific than a typical commercial SaaS privacy policy. There is no central server operated by the maintainers that receives your code, prompts, or usage data through the ordinary operation of the software.
1.1 Who this policy covers
Individuals who download, install, and run the Arch software locally.
Contributors who interact with the project's public repository, issue tracker, or discussion forums.
Anyone who uses optional, clearly-labeled hosted or cloud features, if and when the project offers them.
1.2 Who this policy does not cover
This policy does not cover the privacy practices of third-party large language model (LLM) providers, cloud services, package registries, or other external tools you choose to connect to Arch. If you configure Arch to call an external API (for example, a hosted model provider), your use of that service is governed by that provider's own privacy policy and terms, not this one. See Section 6.
2. Core Privacy Principle: Local-First, No Telemetry by Default
Arch is designed to run fully on your local machine. In its default configuration:
No source code, prompts, file contents, or command output is transmitted to any server operated by the maintainers.
No usage analytics, crash reports, or telemetry are collected or transmitted automatically.
No account, sign-up, or network connection is required to install or run the software.
Indexing, retrieval (RAG-based context lookup), multi-file editing, and verification loops all execute locally, using local compute and local storage.
This means that, unless you explicitly enable an optional feature that requires network access (such as connecting to a third-party model API), Arch has no visibility into your code, your prompts, or how you use the tool.
2.1 What “local-first” means in practice
3. Data Processed Locally
Although Arch does not transmit data externally by default, it does read, generate, and temporarily store certain data locally in order to function. This section describes that processing for transparency, even though it never leaves your machine under default settings.
3.1 Codebase content
To provide multi-file editing and context-aware suggestions, Arch reads files within the project directory you point it to, and may build a local semantic index (via RAG-based indexing) to retrieve relevant context efficiently. This index is stored locally, typically alongside the project or in a local cache directory, and is used only to serve your own queries.
3.2 Prompts and conversation history
Your prompts, the agent's responses, and any intermediate reasoning or tool-call records are kept in local session state to support features like session portability (resuming a session later) and verification loops (checking prior steps before proceeding). You can typically clear this history by deleting the relevant local cache or session files — see your local configuration documentation for the exact path.
3.3 Verification and execution data
Where Arch runs code, tests, or commands as part of a verification loop, the output of those commands is processed locally to inform the agent's next steps. This output is not transmitted anywhere by the core software.
3.4 Local logs
If you enable logging for debugging purposes, logs are written to your local disk only. You control their retention and deletion. The project does not include any built-in mechanism to upload logs to the maintainers.
4. Data the Maintainers Receive
As maintainers of an open-source project, we do receive limited data — but only through channels you choose to use, separate from running the software itself.
4.1 Repository and issue tracker activity
If you open an issue, submit a pull request, comment, or otherwise interact with Arch's public repository (hosted at github.com/Wizzit-254/Arch-Assistant-Repo), the repository host (e.g. GitHub) collects account and activity data under its own privacy policy. Anything you post publicly — including code snippets, logs, or screenshots you choose to paste into an issue — is visible to the public and is your responsibility to redact of sensitive information before posting.
4.2 Direct communication
If you email us or otherwise contact us directly at trevorkisingu@gmail.com, we receive whatever information you choose to include in that communication (e.g. your email address, and the content of your message). We use this only to respond to you and to improve the project.
4.3 Package registry / download statistics
If Arch is distributed via a package manager or registry (such as npm, PyPI, or a container registry), that registry may independently log download counts, IP addresses, or similar technical metadata as part of its own infrastructure operations. The maintainers typically only see aggregate, anonymized statistics (e.g. total download counts), not individual user identities, unless the registry's own policy states otherwise.
4.4 Contributions and commit metadata
If you contribute code, your commits will include the name and email address associated with your git configuration, per standard version control practice. This becomes part of the public project history.
5. Telemetry and Analytics
Arch does not include any telemetry, analytics SDK, or "phone home" behavior in its default build. There is no hidden network call that reports usage statistics, feature usage, or error data to the maintainers.
5.1 Update checks
If Arch checks for new versions of itself, this may involve a network request to the repository host or package registry to compare version numbers. This request follows standard HTTP request patterns (which can include your IP address, as with any web request) but does not include any information about your code, prompts, or usage. You can typically disable automatic update checks in local settings.
6. Third-Party LLM Providers and Integrations
Arch is model-agnostic and can be configured to work with local models or external, third-party model APIs, depending on how you set it up.
6.1 Local models
If you configure Arch to use a model that runs entirely on your machine (e.g. via a local inference server), no prompt or code data leaves your device as a result of model inference.
6.2 External model APIs
If you configure Arch to call an external, third-party LLM API, the prompts and relevant code context you send will be transmitted to that third party according to your configuration. In that case:
The third-party provider's own privacy policy and data retention practices govern that data, not this document.
Arch transmits only what is needed to serve your request (e.g. the active prompt and relevant context) — it does not upload your entire codebase or session history unless your configuration or the model's context window causes it to be included.
You are responsible for reviewing the terms of any third-party provider you choose to connect, especially before sending proprietary, confidential, or regulated code or data through it.
We recommend consulting the documentation for any model provider you configure (for example, its data retention window, whether it trains on submitted data, and regional data handling) before use.
6.3 Other integrations
Arch may support optional integrations with other developer tools (e.g. version control hosts, package registries, or MCP-compatible servers). Each integration only activates data flow to that specific third party if and when you explicitly configure it. Review the third party's privacy policy before enabling any such integration.
7. Data Security
Because Arch operates primarily on your local machine, the security of your data is largely governed by the security of your own device and any third-party services you choose to connect. That said, the project follows these practices:
Local data (session history, indexes, config) is stored using your operating system's standard file permissions; the software does not apply additional encryption by default unless noted in the documentation.
The project does not operate a central database of user data, which eliminates an entire category of breach risk associated with hosted services.
Dependencies are reviewed and updated as part of ongoing maintenance; users are encouraged to keep their installation up to date.
If you connect third-party APIs, credentials (e.g. API keys) are stored locally per your configuration — treat them with the same care as any other secret, and avoid committing them to version control.
8. Your Rights and Choices
Because Arch does not collect personal data through the software itself, most traditional data-subject rights (access, deletion, portability) are naturally satisfied by the fact that your data never leaves your device. You retain full control over it at all times. Specifically, you can:
Delete any local session history, index files, or logs at any time by removing them from your filesystem — see the project documentation for exact file locations.
Disable network-dependent features (such as update checks or third-party model integrations) at any time via local configuration.
Choose which, if any, third-party model provider to use, or run entirely offline with a local model.
Fork, audit, or modify the source code yourself, since the project is open-source — you can verify these privacy claims directly by reading the code.
For data you've shared with third parties (a repository host, a model API provider, or via direct email to us), your rights are governed by that party's own privacy policy and applicable law (e.g. GDPR, CCPA), and you should contact that party directly to exercise those rights.
9. Children's Privacy
Arch is a developer tool not directed at children, and we do not knowingly collect personal information from children through the software itself, consistent with its local-first, no-telemetry design. If you believe a child has provided personal information to us through direct contact (e.g. email or the issue tracker), please contact us at trevorkisingu@gmail.com so we can address it.
10. Open-Source License and Transparency
Arch is distributed under the MIT License. Because the source code is publicly available at github.com/Wizzit-254/Arch-Assistant-Repo, you — or any third party — can independently verify the data-handling claims made in this policy by inspecting the code directly. We encourage community review and welcome issues or pull requests flagging any behavior that is inconsistent with this policy.
11. Changes to This Policy
We may update this privacy policy as the project evolves — for example, if new optional features, integrations, or hosted services are introduced. Material changes will be noted in the project's changelog or release notes, and the "Effective date" at the top of this document will be updated accordingly. We encourage users to review this policy periodically, particularly before enabling new features that involve network access.
11.1 Version history
12. Contact Us
If you have questions about this privacy policy, or about how Arch handles data, please reach out through: