Unified employee sign-in
Members sign in through your IdP (Okta, Azure AD, Google Workspace, JumpCloud, or any OIDC-compliant). No per-person API key configuration. New hires get day-one access; departures are SSO-revoked instantly.
Enterprise · DOSIA Cloud
DOSIA Cloud Enterprise is ByteSpike's team tier — SSO sign-in for every member, model permission templates (global / China presets), per-user usage attribution, and a contractual SLA. Built on the same gateway you can try free; deployable inside your VPC if compliance requires it.
Why team admins choose Cloud Enterprise
Personal accounts work for a developer. Teams hit a wall: per-member keys, fragmented billing, inconsistent model access. Cloud Enterprise removes that wall.
Members sign in through your IdP (Okta, Azure AD, Google Workspace, JumpCloud, or any OIDC-compliant). No per-person API key configuration. New hires get day-one access; departures are SSO-revoked instantly.
One-click 'Global edition' (Claude / GPT / Gemini / DeepSeek V4) or 'China edition' (DeepSeek / Doubao / GLM / Kimi / MiniMax) presets. Admin assigns templates per team; members get exactly the model surface they're authorized for.
Every gateway call attributed to the org member who made it, with per-user monthly budget caps and admin alerts at 80% / 100%. Finance gets a clean per-employee bill; engineering gets visibility into who spent what on which model.
One-click presets
Permission templates ship with two curated presets so the admin doesn't pick 30 models on day one. Apply Global or China with one click, then refine per team — your members only see what you've turned on for them.
Teams operating internationally or with English-first workloads.
Claude · GPT-5 · Gemini · DeepSeek V4
Admin Console → Templates → Apply Global preset
Teams with PRC compliance or PII-sensitive workloads.
DeepSeek · Doubao · GLM · Kimi · MiniMax
Admin Console → Templates → Apply China preset
Custom templates start from either preset, so you mix and match — pick the Global core, swap in DeepSeek Flash for cheap text, drop the image lineup, save as 'engineering-pack'. Every member assigned to that template sees exactly that subset, automatically intersected with your account's allowed model pool.
Try presets in the consoleWho uses it
DOSIA Cloud is built for cross-functional teams that all need AI on their daily work but don't want to wire up keys, base URLs, and provider accounts themselves.
Campaign drafting, ad copy variations, brief generation. Template-pinned to text models with image-gen permission for asset preview.
Long-form writing, editorial review, multi-language translation. Heavy on Claude / GPT-5; cache_control on shared system prompts.
SOPs, internal-doc QA, meeting summaries, decision logs. Reasoner models for analytical work; cheap models for routing.
Ad performance analysis, A/B-test interpretation, audience segmentation. Long-context models for raw data; web search for competitive intel.
Prospect research, outreach drafting, call summarization, CRM enrichment. Permission template excludes image / video to keep costs predictable.
Policy QA, onboarding assistant, internal communications drafting. Highest privacy tier — locked to the China-region deployment for PII-sensitive workloads.
How it fits together
DOSIA Cloud is the team-management and policy layer. The ByteSpike Gateway is the unified API and billing layer underneath. Members never handle model credentials; admins manage policy, not plumbing.
DOSIA Cloud · team management & policy
Member SSO · template permissions · per-user attribution · budget caps
ByteSpike Gateway · unified API & billing
Anthropic Messages · OpenAI Chat / Responses · multimodal
Model families (Anthropic / OpenAI / Gemini / DeepSeek / Moonshot / ...)
Architecture diagram lands here once the DOSIA Cloud Console screenshots are published.
Security & compliance
Everything an enterprise security review asks for, with the option to self-host the gateway when policy requires data never leave your perimeter.
Okta, Azure AD, Google Workspace, JumpCloud, or any OIDC IdP. SCIM provisioning for org-member sync. Per-org role mapping (admin / developer / viewer) with per-key scope (quota · rate-limit · IP allowlist) inheritance.
Every gateway call recorded with model · key · org-member · outcome · cost. Queryable via API or exported to S3 / GCS. Retention configurable up to 7 years for compliance regimes.
99.95% uptime on the gateway, p99 latency commitments per model family, monthly credits issued automatically when targets miss. The SLA lives in your contract, not on a marketing page.
Same wire protocol as bytespike.ai, runs in your VPC. Docker Compose for single-node; Kubernetes Helm chart for clustered. Air-gapped deployments with quarterly model-weight refreshes available on request.
Async tasks (video generation, batch jobs) post completion webhooks to your endpoint with HMAC-signed payloads. Configurable retry budget, dead-letter queue, replay-from-timestamp. No polling required.
Assigned at contract signing. Slack Connect or Teams channel for real-time issues. 4-hour SLA on Sev-1, next-business-day on Sev-2. Quarterly architecture reviews included.
Pricing
Personal is for one developer; Enterprise is for the team. Both consume the same gateway underneath — the difference is the team-admin layer above it.
Pay-as-you-go
ByteSpike credits · no minimum
Contact sales
Per-user quota · custom contract · annual billing
Contact sales
Tell us your team size, expected model mix, security requirements. We'll come back within 3 business days with a contract draft. No deck demo, no follow-up spam — if we're not the right fit, we'll say so.