Answers, before you ask.
Everything customers most often want to know about Framer, gathered in one place.
Product
What does Framer actually do?+
Framer is an AI platform for customer support. It reads tickets, retrieves knowledge, drafts replies, automates workflows, and reports outcomes — all inside your existing helpdesk or as a standalone workspace.
Is it a chatbot?+
No. Chatbots follow scripts. Framer reasons over your data, cites its sources, and acts across your stack. Agents stay in the loop as much or as little as you want.
Can it replace my helpdesk?+
It runs alongside Zendesk, Intercom, Freshdesk, Salesforce and Front, or as a standalone workspace. Your choice.
Which channels are supported?+
Email, live chat, in‑app messaging, voice, WhatsApp, SMS, social DMs, community forums and custom channels through our SDK.
How many languages does it speak?+
104 languages with automatic detection and bi‑directional translation. Tone and formality are preserved per locale.
Does it write in our brand voice?+
Yes. Style transfer is trained on approved historical replies, help center content and a style guide you can edit at any time.
Can it handle voice calls?+
Real‑time transcription, agent‑assist and post‑call summaries are included. Full autonomous voice is available on Enterprise.
What models power it?+
A blend of frontier LLMs (OpenAI, Anthropic, Google) plus our own routing and retrieval models — abstracted behind one API.
Can we bring our own models?+
Yes — Enterprise supports BYOM with Azure OpenAI, AWS Bedrock, Vertex AI or a private endpoint.
How does it stay accurate?+
Every reply is grounded in retrieved sources with citations, then verified against your policies before it sends.
What about hallucinations?+
Answers with insufficient evidence are escalated instead of guessed. Confidence scores are exposed to agents and workflows.
Does it work offline / on‑prem?+
An air‑gapped deployment with local inference is available for regulated customers on Enterprise.
How does the AI learn from feedback?+
Agent edits, thumbs, CSAT and outcome signals continuously retrain your private retrieval and style layers — never shared models.
Can non‑support teams use it?+
Yes — IT helpdesk, HR, revenue ops and success teams use Framer for internal and external requests alike.
Deployment
How long to go live?+
Most teams see meaningful deflection in week one. Full deployment averages 3 weeks including tuning and integrations.
Do we need engineers?+
No — connect sources in three clicks. Engineering is only needed for custom actions or private deployments.
Can it run in our VPC?+
Yes — enterprise VPC and BYOK deployments are available.
What does onboarding look like?+
Day 1 connect helpdesk and knowledge, day 2 train on 90 days of tickets, day 3 shadow mode, week 2 go‑live with guardrails.
Do you offer a sandbox?+
Every workspace comes with an isolated sandbox for testing prompts, workflows and integrations against synthetic traffic.
How do you migrate from a legacy vendor?+
We import tickets, macros, tags, users and reporting from Zendesk, Intercom, Salesforce, Freshdesk and Front with zero downtime.
Do you support hybrid clouds?+
Yes — control plane on our cloud, data plane in yours (AWS, GCP, Azure) with mTLS between the two.
What about staging environments?+
Unlimited environments on Growth and Enterprise, each with its own connectors, secrets and models.
Can we deploy region‑locked?+
Yes — pin data, models and logs to EU, US, UK, Canada, Australia, Japan or India.
How is downtime handled?+
Active/active across three availability zones with automatic failover. 99.99% uptime on Enterprise.
Do you provide implementation partners?+
A certified partner network delivers migrations, custom workflows and change management on request.
What about ongoing tuning?+
Growth and Enterprise include a dedicated success engineer with monthly reviews and quarterly business reviews.
Can we roll back a change?+
Every workflow, prompt and knowledge version is diffable and one‑click revertible.
Is there a CLI or IaC?+
Yes — Terraform provider, CLI and Git‑ops so your workspace is a versioned artifact, not a screen recording.
Do you offer white‑glove launch?+
Enterprise includes a launch pod: solution architect, prompt engineer, data engineer and CSM for the first 90 days.
Security
Do you train on our data?+
Never on shared models. Your data trains only your private tenancy, and only when you enable it.
Certifications?+
SOC 2 Type II, ISO 27001, HIPAA, GDPR, PCI DSS. FedRAMP and C5 in progress.
Data residency?+
EU, US, or APAC. Data never leaves the chosen region.
How is data encrypted?+
AES‑256 at rest, TLS 1.3 in transit, per‑tenant KMS keys with optional customer‑managed keys.
Do you support SSO and SCIM?+
SAML 2.0, OIDC and SCIM 2.0 with Okta, Azure AD, Google Workspace, OneLogin, JumpCloud and Ping.
How is PII handled?+
Automatic detection and redaction for 40+ PII types before prompts hit any model, with per‑field retention policies.
What audit trails exist?+
Every AI action, prompt, tool call, human edit and configuration change is logged with immutable timestamps and streamed to your SIEM.
Do you run pen tests?+
Third‑party pen tests twice a year plus a continuous public bug bounty on HackerOne.
What's your incident response?+
24/7 on‑call, 15‑minute SEV‑1 acknowledgement, customer notification within one hour of confirmation.
How do you handle sub‑processors?+
A public sub‑processor list with 30‑day advance notice of changes and per‑customer opt‑outs where legally required.
Is data ever used cross‑tenant?+
Never. Physical, logical and cryptographic isolation with per‑tenant vector indexes.
What about model access?+
Prompt/response logs to model providers are opted out globally. Enterprise adds zero‑retention agreements with each provider.
Can we self‑host inference?+
Yes — bring your own model endpoint or run our reference stack in your VPC with GPU or CPU inference.
Do you support customer‑managed keys?+
AWS KMS, GCP KMS and Azure Key Vault, with rotation policies you control.
How long is data retained?+
Configurable from 24 hours to 7 years per data type; hard deletes are cryptographic, not tombstoned.
Pricing
How much does one credit cost?+
One credit is a flat $0.10. The minimum order is 50 credits ($5).
Do you have plans or seats?+
No. There are no plans, no seats and no trials — everyone uses one credit-based currency. Team members are unlimited on every workspace.
What can I spend credits on?+
Every Platform module (AI Assistant, Knowledge Search, Ticket Intelligence, Reply Generator, Workflow Automation, Help Center AI, Analytics, Integrations, Security) and every Solution pack (Customer Support, IT, SaaS, E-commerce, Healthcare, Finance, Enterprise, Small Business).
How is usage metered?+
Each module has a transparent per-event rate (e.g. 5 credits per resolved ticket, 1 credit per knowledge query). Every debit shows up on your credit ledger in real time.
Do credits expire?+
No. Credits stay on your workspace as long as it is active.
What are bonus credits?+
Orders of 500+ credits get +5%, 2,000+ get +10%, and 10,000+ get +15% bonus credits added on top of your purchase.
Do you offer promo codes?+
Yes — try WELCOME10 for 10% off your first order. Enterprise and multi-year customers get custom codes.
How do Solution packs work?+
Each Solution is a one-time activation per workspace — pay once in credits and it stays installed with all its tuned models, workflows and integrations.
Can we cap spend?+
Yes — set hard caps, soft warnings, and per-module budgets in Settings → Billing to prevent surprise usage.
Can I turn modules off?+
Any Platform module can be paused from the dashboard. Credits stop being consumed the moment it is off.
How do we pay?+
Bank transfer is available today: place your order and we email the IBAN/SWIFT and reference. Card, PayPal and Stripe are coming soon.
What currencies are supported?+
Credit prices are billed in USD; invoices can be issued in EUR, GBP, CAD, AUD, JPY, SGD and INR.
Can we pay by invoice?+
Yes, for orders of 2,000+ credits — NET-30 by default, NET-60 on request.
Refund policy?+
Unused credits are refundable within 30 days of purchase, minus payment processing fees.
Where do I buy more?+
Any time from Settings → Billing → Buy credits inside your dashboard.
Give your support team an unfair advantage.
Buy 50 credits. Deploy in a day. Measure lift in your first week.