Question 1

How fast is the API? Will it add latency to our LLM requests?

Accepted Answer

promptward is optimised for sub-50ms latency (p95). Most requests complete in 10–20ms. We sit between your application and the LLM provider, processing in parallel with your request. For context: OpenAI API latency is typically 100–500ms. Our overhead is negligible. We also support batching and caching for high-volume scenarios.

Question 2

Which LLMs do you support? What about custom/self-hosted models?

Accepted Answer

We support OpenAI, Anthropic, Google, Cohere, Mistral, and any LLM accessible via REST API (including self-hosted, Ollama, vLLM). If your provider has an API, we can sit in front of it. We're provider-agnostic by design — no lock-in.

Question 3

How much engineering effort to integrate? Will we need to refactor?

Accepted Answer

Three lines of code. No refactoring required. If you're already calling an LLM API, you replace that call with a promptward call. It's a drop-in middleware. Integration typically takes 1–2 hours including testing. No vendor lock-in — if you decide to remove us, you revert to direct API calls.

Question 4

What about cost? Does this add per-token fees?

Accepted Answer

Transparent, usage-based pricing. Early signups get locked-in pricing: $99–$499/mo depending on tier (see pricing), regardless of token volume. No per-token surcharge. We pass provider costs through directly; we only charge for governance and audit.

Question 5

Can we see the code? Is it open source?

Accepted Answer

The core API is closed-source (that's where the value is). But we provide open-source SDKs (Python, TypeScript, Go) so you can audit the integration code. Webhooks and event streaming are documented publicly. We're transparent about what we log and how we handle data.

Question 6

Can we export our audit logs? Do you provide compliance reports?

Accepted Answer

Yes. Audit logs can be exported in real time via API or downloaded as CSV. We provide: sanitised data reports, redaction logs, threat detection reports, and user access logs. All queryable by date range, user, or asset. For regulated customers, we'll work with your compliance team to format reports for your auditors.

Question 7

What data do you actually store? How long do you keep it?

Accepted Answer

We store: request metadata (timestamp, user ID, LLM provider, tokens used), redaction results (what was redacted, why), threat detection results (injection attempts detected), and audit logs. We do NOT store plaintext prompts or responses by default — only redacted versions and metadata. Retention is configurable (default: 30 days for logs, 7 years for audit trail). You control the retention policy.

Question 8

Are you SOC 2 / ISO 27001 certified? What about regulatory audits?

Accepted Answer

SOC 2 Type I certification is in progress (expected Month 6–12). We're building toward ISO 27001. For now: we're hosted on AWS with encrypted data at rest and in transit, regular security audits, and comprehensive data handling documentation. If you need SOC 2 before launch, we can prioritise your audit; early signups get priority access to compliance certifications.

Question 9

GDPR, CCPA, Australian Privacy Act — are you compliant?

Accepted Answer

Yes to all three. GDPR: we support data access, rectification, and deletion rights. CCPA: opt-in consent, deletion requests, no sale of data. Australian Privacy Act: data residency options (AU-hosted on request), Australian Privacy Principles aligned. We don't sell data, don't share with third parties, and maintain encryption by default.

Question 10

Can you help our team pass a compliance review? Do you provide documentation?

Accepted Answer

Yes. We provide: architecture documentation, data flow diagrams, security assessment results, compliance checklists, and templated compliance questionnaires. We can also do a live walk-through with your compliance or security team (subject to availability). Early signups get white-glove compliance support included.

Question 11

What's the pricing? Do you have a free tier?

Accepted Answer

Free tier: up to 10K API calls/month (evaluate promptward, build a POC). Startup tier: $99/mo (unlimited calls, audit logs, basic support). Growth tier: $299/mo (priority support, custom retention, API SLA). Enterprise: custom pricing (white-glove support, compliance assistance, SLA negotiation). Early signups (from waitlist) get 20% lifetime discount on any tier + 90-day free trial.

Question 12

What do I get as an early signup? Is there a discount?

Accepted Answer

Yes. Early waitlist members (first 100) get: (1) 20% lifetime discount on your chosen tier, (2) 90-day free trial (no credit card required), (3) Priority onboarding (personal engineer assigned), (4) Early access to new features (beta program), (5) Founding team badge on launch (if publicly featured). Essentially: lock in lower pricing permanently + get VIP treatment during early phase.

Question 13

Can we trial before committing? How does the free trial work?

Accepted Answer

90-day free trial for all waitlist members — no credit card required. You get full API access, all logs, admin dashboard. At day 90, you decide: upgrade to a paid plan (discount applied) or cancel. No strings attached. If you hit the free tier ceiling (10K calls/mo), we'll email you; you can upgrade anytime.

Question 14

Can promptward work with our existing tools? (Slack, DataDog, etc.)

Accepted Answer

Yes. We provide webhooks so you can send events to Slack, DataDog, or any webhook consumer. Example: threat detected → Slack notification to security team. We also have REST API + SDKs, so custom integrations are straightforward. And: we're building integrations with common tools (more details at launch).

Question 15

Will promptward work with our LLM gateway / observability stack?

Accepted Answer

Yes. We're middleware — we sit between your application and your LLM provider. If you're using LiteLLM, Portkey, or any gateway, promptward can wrap around it. If you're using DataDog / Arize for observability, our events integrate cleanly (via webhooks or direct API). We're designed to be part of your existing stack, not replace it.

Sanitise every prompt.
Prove it with the record.

If you're shipping AI, this probably sounds familiar.

Compliance won't sign off

Security is stitched together

Good engineers, wrong problem

One gate. Every prompt passes through it.

Inspect

Redact

Guard

Record

Everything a governance layer should do. Nothing you have to build.

Blocks prompt injection

Redacts PII and secrets

Keeps the record

Works with any LLM

Drops straight in

Compliance-first, AU-built

The questions your team will ask.

Early teams get the founder's deal.

Stop rebuilding
compliance plumbing.

Sanitise every prompt.Prove it with the record.