Terms of Service
Effective Date: October 14, 2026 • Last Updated: October 30, 2026 • Operator: SentioGate Team / sentiogate.app
1. Acceptance of Terms
By accessing or using the SentioGate website, dashboard, REST API or client libraries (the "Service"), you agree to these Terms. If you accept on behalf of an organization, you confirm you have authority to bind it. If you do not agree, do not use the Service.
Minimum age. You must be at least 16 years old, and old enough to enter a binding contract where you live, to create an account or use the Service.
2. The Service
SentioGate is an OpenAI-compatible routing gateway. For each request it picks a model from the models you enabled, cost first by task tier, with measured latency and error rates as a secondary signal, and forwards the request with your key.
- Bring your own key. Once you store a provider key in your Vault, every request is sent with your own keys and the Service never uses platform-owned provider keys for your account. You keep a direct relationship with each provider (for example OpenAI, Groq, DeepSeek, Mistral, xAI, OpenRouter or a custom endpoint) and are responsible for following its terms and acceptable-use policies.
- Community pool (beta). During the beta, an account with no provider keys in its Vault can use a free, rate-limited community pool from the dashboard and the API (
model: "auto"or"sentiogate-community"). These requests are not BYOK: SentioGate sends them to Agnes AI using a platform-provided key, and Agnes AI processes them under its own terms, which allow it to use inputs and outputs to improve its service unless an opt-out applies (see the Privacy Policy). Do not send personal, confidential or regulated data through the community pool. Limits are 600 requests per rolling 5 hours per account; requests you choose to queue count against the same allowance. Answers are AI-generated and may be wrong. The pool can be limited, changed or ended at any time. Adding any provider key to your Vault switches your account to BYOK and turns the pool off. - No token markup. SentioGate never marks up provider prices. Providers bill you directly; SentioGate fees cover the software and the Service only.
- Response cache. On for non-streaming requests without tools when temperature ≤ 0.3, or when temperature is unset and the prompt is a single-turn factual question (math, multiple choice, short factual Q&A, classification; never creative prompts), or when you opt in (
X-SentioGate-Cache: forceor"cache": true). Cached responses are kept for 1 hour (10 minutes for unset-temperature factual requests) and may be returned for identical or very similar requests from your account. You can opt out per request (headerX-SentioGate-Cache: offor"cache": false) or flush your cache. - Context compression. On by default. Lossless clean-up (duplicate messages and extra whitespace; system prompts, code and tool calls untouched) always runs. Conversations above about 16k tokens also get older turns summarized by your cheapest capable model, using your key, and that summarization call is billed to you by the provider. Opt out per request with
compress: offorlossless. - Semantic cache embeddings use Cloudflare Workers AI when available; otherwise your OpenAI key (text-embedding-3-small), billed to you by OpenAI.
- Quality grading. When quality grading is on, a small sample (about 2%, at most one per minute per account) of auto-routed prompts and answers with a checkable answer is sent once more to your own provider on your own key, using your cheapest other enabled model, for a correct/incorrect verdict that improves routing. That grading call is billed to you by the provider. Only aggregate counts per model and task type are stored, no prompt or answer text.
- Discover. Files you upload for a savings analysis are processed in memory and not stored. Savings figures are estimates from your data and public price lists, not guarantees.
3. Availability, Fallback & Rate Limits
When a model errors or is rate-limited, auto routing retries the next-cheapest qualifying model, and it skips models whose recent measured success rate or latency is poor. Providers still enforce their own limits, credits and outages, and we do not guarantee that any request succeeds. SentioGate is not liable for delays, failed or degraded responses, or interruptions caused by providers, model deprecations, account suspensions or rate limits.
You agree to indemnify SentioGate and its operators against claims arising from your use of providers through your keys, or from content you send through the community pool, including runaway agent loops and acceptable-use violations.
4. Data & Processing Locations
The Service runs on Cloudflare Workers and stores data in Cloudflare D1. Requests are processed in the Cloudflare location that serves them and then sent to the provider you route to, wherever it processes data. Community pool requests are sent to Agnes AI, which states that it processes data in Singapore, the United States and other countries. Your data may therefore be processed outside the country you live in. By using the Service you authorize these transfers. The Privacy Policy lists what is stored and for how long. No data residency guarantee is provided.
5. Security & Key Vault
Vault keys are encrypted at rest using a key derived from a server-side secret and a random salt per key. They are decrypted only in memory to call the provider. Prompts and completions are not written to logs; the only prompt-derived data stored is the 1-hour response cache described above and, for community-pool requests you choose to queue, the queued request and its answer for up to 3 days. Keep your gateway keys secret: anyone holding one can send requests that use your provider keys.
6. Proprietary Software & License
The SentioGate hosted service and platform are proprietary, closed-source software. You get a limited, non-exclusive, non-transferable right to use the Service. You may not copy, resell or redistribute the Service, or reverse-engineer it except where the law allows.
Exception: the JavaScript/TypeScript client SDK (@sentiogate/sdk) and Python client SDK (sentiogate-python) are open-source SDK licensed under the Apache License 2.0. Their licenses govern your use of the SDK code; these Terms govern your use of the Service they connect to.
7. Fees & Plans
- Plan and license fees are shown on the Pricing page. All accounts currently receive a 90-Day Free Beta envelope: 25,000 routed requests per calendar month (UTC) at $0 platform fee, enforced as a soft monthly cap: once the month's count is reached, requests return an HTTP 429 error until the next month (counts update about once a minute, so a few requests may pass the cap). Requests on your own keys are billed by each provider; the community pool is free during the beta.
- Fair use. The free beta envelope and the community pool are for your own development and evaluation. We may throttle, suspend or close accounts that abuse them, for example by creating several accounts to multiply allowances, reselling or sharing access, or trying to get around limits.
- Plans can include request or provider limits, which we may enforce with rate limiting.
- You are responsible for applicable taxes (GST, VAT or similar).
8. Disclaimer of Warranties & Limitation of Liability
The Service is provided "as is" and "as available", without warranties of any kind, including merchantability, fitness for a particular purpose and non-infringement. Routing decisions, cost estimates and savings reports are estimates. The Service, including the community pool, is a beta: features can change or be removed without notice, and it may be unavailable or lose queued requests.
SentioGate is not liable for indirect, incidental, special, consequential or punitive damages, or for lost profits or data. Total liability is limited to the fees you paid SentioGate in the twelve (12) months before the claim.
9. Governing Law
These Terms are governed by the laws of India, without regard to conflict-of-law rules. Disputes are subject to the exclusive jurisdiction of the competent courts in India.
10. Contact
- Technical Support: support@sentiogate.app
- Enterprise & Licensing: sales@sentiogate.app
- Security Disclosures: security@sentiogate.app