Startup Engineering Pod (Scenario)
~58% blend savings
Shifting 60% of routine summarization calls to lightweight models cuts blended token spend with 0% BYOK markup.
One privacy and cost-routing shield across your Mobile, Desktop, and Production APIs—scrambling sensitive data and routing every prompt to the right model before anything leaves your boundary.
Live pre-egress trace
Scrubbing, smart routing, and spend caps.
Sit Spendplane in front of IDEs, Python/Node SDKs, n8n/Make/Zapier automations, and MCP tool servers in 15 seconds—with scoped keys and pre-egress PII scrubbing per tool.
IDEs & Coding
App Frameworks & SDKs
No-Code & Ops Middleware
Enterprise & Agent Infra
Override OpenAI Base URL with a scoped project key and safe model alias.
2. Choose default policy preset
Pre-egress PII and secret scrubbing enabled before any external model call.
IDEs & Coding Agents
IDE Settings (OpenAI Base URL Override)
# Settings -> Models -> Override OpenAI Base URL
Base URL : https://api.spendplane.com/v1/proxy/projects/proj_demo_9f2a/v1
API Key : sp_live_your_scoped_key
Model : auto
Preset : strict_privacySub-5ms pre-egress redaction tokenizes emails, API keys, and sensitive records ([EMAIL_1], [SECRET_1]) before prompts leave your boundary—with zero prompt body storage.
Connect your own OpenAI, Anthropic, local Ollama, or private GPU endpoints (RunPod, vLLM, cloud clusters). Split traffic across multiple models in parallel to boost throughput, fail over automatically, and cut blended token costs—without changing your app code.
Explore Smart RouterIndividuals and startups use Spendplane Hosted Gateway + AI Safe on-device. Enterprise security teams deploy Spendplane inside their own AWS, GCP, or Azure VPC with SAML SSO, SCIM, SIEM log streaming, and signed DPAs.
Traditional VPNs mask your IP address while sending raw prompt text straight to AI providers. Spendplane's Data VPN Firewall protects your data points before egress—across mobile devices, desktops, and production APIs.
Mobile (iOS & Android)
Protected Mobile AI Workspace
Use mobile AI safely inside a protected workspace with on-device pre-egress scrubbing, real-time budget alerts, and instant key/session kill-switches.
Before Spendplane
surface: AI Safe Mobile Workspace (iOS & Android)
prompt: "Summarize escalation from alex@acme.com"
secret: "card ending 4421 / key sk-live-94a8..."
After Pre-Egress Shield
engine: AI Safe Mobile (Pre-Egress Scrubbing)
egress: "Summarize escalation from [EMAIL_1]"
controls: live spend alert OK • kill-switch ready
Pre-Build Tool • Spendplane Planner (Free)
Cost & Data-Path Simulator — model token spend, data sensitivity, and routing rules before writing a line of code.
The governance tunnel
Spendplane captures the request, applies policy, picks the cheapest viable lane, and shows the live decision instead of hiding it behind the agent.
Sub-5ms pre-egress redaction tokenizes PII and API secrets ([EMAIL_1], [SECRET_1]) before prompts leave your boundary, while enforcing workspace spend caps with zero prompt retention.
Not every prompt needs a frontier reasoning model. Spendplane dynamically routes your payload to the cheapest capable model based on task complexity, slashing your cloud bill without sacrificing quality.
Watch the full lifecycle in real-time. Once the LLM responds with scrubbed tokens, Spendplane securely re-maps them to their original values and delivers a complete, private response to your app - zero PII ever leaked.
_
The biggest blocker to enterprise AI is sensitive data. Spendplane detects and tokenizes PII and secrets from outbound requests before they ever reach a third-party LLM.
Organizations handling sensitive citizen or customer identifiers enforce strict pre-egress redaction and regional routing.
Auto-detected Entities
Real-time Interception
Please analyze the tax history for citizen QQ 12 34 56 C holding passport P83726190.
Please analyze the tax history for citizen [UK_NINO_1] holding passport [PASSPORT_1].
Illustrative 1M-token blend with BYOK pass-through (0% token markup)
Math: Shifting 60% of routine summarization or classification prompts from a $15/1M frontier lane to a $0.50/1M lightweight lane yields ($15 × 0.4) + ($0.50 × 0.6) = $6.30 (~58% lower blended spend).
Model your expected monthly token volume, compare cloud providers against local Ollama or GPU runners, and preview how routing rules affect blended API costs before shipping.
Every Spendplane plan uses BYOK pass-through with 0% token markup—you pay your model provider directly while Spendplane enforces redaction, routing, and spend caps.
How teams combine AI Safe on-device protection with Spendplane Gateway routing and budget enforcement.
Startup Engineering Pod (Scenario)
~58% blend savings
Shifting 60% of routine summarization calls to lightweight models cuts blended token spend with 0% BYOK markup.
Product & Support Team (Scenario)
Sub-5ms redaction
Tokenizes customer emails and API keys pre-egress across desktop browsers, IDEs, and production support bots.
Regulated Enterprise VPC (Scenario)
0 prompt bytes stored
Self-hosted VPC gateway streams metadata-only audit events to Datadog/Splunk without retaining prompt payloads.
Developers and knowledge workers routinely paste API keys, customer emails, and internal identifiers into AI tools. Spendplane inspects every outbound prompt across Mobile, Desktop, and API Gateway before it reaches external models.
Start free, upgrade to Pro for unlimited Mobile, Desktop & API Gateway protection, or scale across Teams and Enterprise VPCs.
Planner & Basic Shield ($0 / forever — no credit card, no auto-trial).
Unlimited Planner & Cost Estimator runs
AI Safe Mobile & Desktop (Basic): local secret & API-key redaction (up to 500 prompts/mo)
Single personal workspace
No credit card, no auto-trial.
Unlimited Mobile, Desktop & Browser AI Safe plus API Gateway ($15.83/mo billed annually).
Unlimited Mobile + Desktop AI Safe (full PII scrubbing & scrambling)
API Gateway (BASE_URL): up to 25,000 protected API requests/mo
Smart Model Routing (BYOK, 0% token markup) + local Ollama/GPU spillover
$15.83/mo billed annually (Save 17%).
Per-seat workspace (min 3 seats, $40.83/seat/mo billed annually) across Mobile, Desktop & APIs.
+ Everything in Pro across all team members' Desktop, Mobile & APIs
150,000 protected API requests/mo + shared workspace spend caps
Central redaction policies (enforce masking rules across all devs)
$40.83/seat/mo billed annually (Save 17%).
Annual contract (starting ~$1,500+/mo) with private VPC deployment & SSO.
Unlimited API requests & custom rate limits
Self-Hosted / Private VPC Gateway (zero third-party cloud transit)
SAML/OIDC SSO, SCIM provisioning & RBAC
Annual contract (starting ~$1,500+/mo).
How Spendplane masks sensitive payload data points—rather than just IP addresses—across Mobile, Desktop, and Production APIs.
Traditional VPNs hide where your traffic comes from by masking your IP address—but they send 100% of your raw prompt text straight to the AI provider. Spendplane routes your AI traffic through a dedicated Data VPN Firewall: instead of masking your IP, it inspects and scrambles sensitive data points, API keys, and PII inside the tunnel before anything reaches an external model.
Spendplane operates as a Data VPN Firewall across every surface: on Mobile, it provides a protected mobile AI workspace with live budget alerts and kill-switches; on Desktop, it protects browser AI chats and IDEs; and for backend applications (Spendplane Gateway), it enforces pre-egress redaction, routing, and spend caps in 60 seconds—with optional self-hosted VPC deployment for Enterprise.
Minimal. Sub-5ms pre-egress redaction and budget checks run with negligible overhead, while automatic failover and fast-lane model routing often improve end-to-end response times.
Create a workspace key, point your OpenAI or Anthropic SDK base URL to Spendplane, and enable your redaction and budget rules in the dashboard. No application rewrite required.
Never. Spendplane scrubs PII transiently and never stores prompt bodies or model completions—only structured metadata (timestamp, model, token counts, and redacted entity categories) retained for 7 days (Free), 30 days (Pro), or 90 days (Teams).
Yes. You can toggle hard daily or monthly spend caps and instant kill-switches for every workspace or API key, plus soft alerts before budgets run out.
Yes. Via the Spendplane CLI or desktop client, you can bridge Spendplane to an active local runner (Ollama, LM Studio, or vLLM) to execute routine or low-complexity prompts locally while bursting heavier reasoning workloads to approved cloud models.
Pro ($19/mo, or $15.83/mo billed annually) is designed for individuals needing unlimited Mobile & Desktop AI Safe plus 25,000 protected API Gateway requests/mo and 30-day audit metadata. Enterprise adds self-hosted VPC deployment, SAML/OIDC SSO & SCIM, SIEM log streaming, signed DPA/BAA, and custom SLAs.