Terms of Service

By accessing or using daedalusdevelopmentgroup.com or any service offered by Daedalus Development Group (DDG) — including but not limited to Logos fine-tuning, the Priscus P1 research waitlist, Erkos on-prem deployments, and the rest of our portfolio — you agree to these Terms of Service ("Terms"). If you do not agree, do not use the services.

These Terms incorporate by reference our Privacy Policy and our Refund Policy. They form a single agreement between you and DDG.

DDG provides several services described publicly on daedalusdevelopmentgroup.com. Notably:

• Logos — paid fine-tuning of open-weight language models on customer-supplied training data. Four tiers (Starter, Studio, Atelier, Custom) with optional model surcharges. The first three are prepaid one-time fees via PayPal, card (when enabled), or NOWPayments crypto. The Custom tier is email-quoted for off-catalog or full-parameter engagements and is billed per-engagement (not via on-site checkout).
• Priscus P1 waitlist — free signup for notifications about a research-stage language model.
• Other portfolio products (Innova, Dominus Node, Lucrum Protocol, Solidus Network, JaCode, opentine, Project Lavalamp) — described on the site for informational and partnership purposes.

Service descriptions, prices, and availability may change at any time. The version of the service offered at the time of your order is the version we will deliver against.

You must be at least 18 years old (or the age of legal majority in your jurisdiction) and capable of entering into a binding contract. If you are using the services on behalf of an organization, you represent that you have authority to bind that organization to these Terms.

You agree NOT to use DDG services for:

• Training data that infringes third-party intellectual property (copyrighted text, images, audio you do not have rights to use).
• Training data containing personally identifiable information (PII) of third parties without their lawful consent or other valid legal basis.
• Producing or facilitating illegal content (CSAM, fraud schemes, weapons or controlled substances tutorials, doxxing, harassment campaigns, etc.).
• Producing models intended for impersonation of real living people, fraudulent identification, or violation of biometric/identity laws.
• Circumventing the export-control restrictions of any base model's license (we may decline to fine-tune models for customers in restricted jurisdictions).
• Any use that violates applicable law in your jurisdiction or in the United States.

We may decline to start, pause, or cancel any job we reasonably believe violates this section. Cancellations under this section are non-refundable.

Your training data is processed strictly per our Privacy Policy: used only to fulfill your order, never used to train other models, and permanently deleted on completion.

• All prices are in United States Dollars (USD) and are due at the time of order.
• Tier base prices and per-model training surcharges are shown transparently in the order wizard. The total displayed at checkout is the total charged.
• Payments are processed by PayPal, Stripe (when enabled), or NOWPayments. We never collect or store card numbers, bank account details, or crypto private keys ourselves.
• Crypto payments are denominated in USD. The customer pays the USD-equivalent amount in the chosen cryptocurrency at the rate locked in by NOWPayments at invoice time.

Training time depends on multiple factors that vary per order: dataset size, base model size and architecture, training technique (LoRA, full-parameter, DPO), number of epochs, add-ons selected, and current GPU availability. We do not commit to a specific turnaround on the order page because doing so would be misleading — small Starter LoRAs can complete in hours, while full-parameter 70B with DPO on 100K examples can take multiple days.

After we receive your training data, we provide a specific timeline estimate at intake and update you if execution conditions change materially. Atelier tier orders receive priority queue placement ahead of Starter and Studio orders.

Service-delivery delays do not entitle you to a refund unless the delay materially exceeds our intake estimate without communication, in which case our Refund Policy governs remedies.

Refund eligibility, cancellation windows, and operator-failure policies are described in our Refund Policy. Once an order is delivered, it is non-refundable.

DDG services are provided on an "AS IS" and "AS AVAILABLE" basis. To the maximum extent permitted by law, we disclaim all warranties, express or implied, including warranties of merchantability, fitness for a particular purpose, and non-infringement.

We do not warrant that any fine-tuned model will achieve a specific accuracy, throughput, or business outcome. Fine-tuning quality depends substantially on the quality and fit of the training data you provide. We deliver an evaluation report so you can assess the model honestly before deploying it.

Open-weight base models (Llama, Qwen, Gemma, Mistral, etc.) come with their own model cards and known limitations. You agree to evaluate any fine-tuned model in your own deployment environment before relying on it for production decisions.

To the maximum extent permitted by law, DDG's total cumulative liability to you for any and all claims arising from or relating to these Terms or the services is limited to the amount you paid DDG for the service that gave rise to the claim, in the twelve (12) months preceding the event.

DDG shall not be liable for any indirect, incidental, special, consequential, or punitive damages, including lost profits, lost data, business interruption, or reputational harm, even if advised of the possibility of such damages.

You agree to indemnify and hold DDG harmless from any claim, demand, or damages (including reasonable attorneys' fees) arising from your training data, your deployment of any model we deliver, or your violation of section 4 (Acceptable Use). This means: if a third party sues us because you uploaded copyrighted training data without rights, you cover our defense costs.

We may suspend or terminate your access to the services at any time for violation of these Terms, for abuse of our infrastructure, or for any other reason at our discretion — though we will refund any prepaid amounts for undelivered work where the termination is initiated by us and is not the result of a violation by you.

You may stop using the services at any time. Termination does not retroactively waive fees for delivered work.

We will treat any non-public information you mark as confidential (or that is clearly confidential by nature) with reasonable care. For Atelier-tier and Erkos engagements, we will sign a mutual NDA on request.

Conversely, certain DDG architecture and methodology — particularly Priscus P1 internals and our Erkos deployment patterns — is held as trade secret. Accessing public marketing materials does not grant rights to internal designs or methodologies; technical briefings under NDA are arranged separately.

These Terms are governed by the laws of the State of Florida, United States, without regard to conflict-of-laws principles. Any dispute arising from these Terms shall be brought exclusively in the state or federal courts located in Florida, and you consent to personal jurisdiction there.

We may update these Terms from time to time. The version in effect at the time of your order applies to that order. Continued use of the site after material updates constitutes acceptance of the new Terms.

Daedalus Development Group
Email: 0xcircuitbreaker@protonmail.com
Web: daedalusdevelopmentgroup.com