Wallet context layer

AI agent context sharing, inside every wallet.

Give Codex, Claude Code, Cursor, and internal agents the same wallet-scoped skills, project context, budget labels, route policy, and usage evidence without copy-pasting prompts.

Reusable skillsPolicy gatewaySpend evidenceNo workspace switch
Wallet context packetknowledge: release-review@approvedroute_policy: wallet-safeusage_report: attached

How it works

AI agent context sharing from the wallet.

Quotaflow packages reusable knowledge, context, budget labels, route policy, and AI spend evidence inside the wallet, then lets agents fetch the right packet at the moment of work.

1

Connect

Add Quotaflow as a custom connector in any Quotaflow-capable agent client.

2

Ask

Describe the task. The agent can request route options, skills, shared context, or spend data.

3

Get

The agent gets the route, context packet, cost estimate, and report trail it needs.

Agents

Works in the agents you already use.

Keep the agents your team already uses. Quotaflow adds shared skills, route policy, and reporting without requiring a new AI employee workspace.

ClaudeClaude CodeCursorVS CodeCodexAny Quotaflow client
managed

Hosted

https://mcp.quotaflow.ai

Point your client at the hosted server. Nothing to install for a first evaluation.

team policy

Gateway

qf mcp connect --workspace team

Run through your Quotaflow gateway when teams need policy, owners, and route reports.

Why shared context

Stop rebuilding context in every thread.

Teams repeat the same instructions, rules, and project knowledge across agents. Quotaflow makes those packets reusable, governed, and measurable.

Ad hoc

Every agent rebuilds context.

{
  "model": "provider/model",
  "context": "pasted manually",
  "workflow": "unknown",
  "budget": "not attached"
}
Look up model, policy, price, and context by hand.
Quotaflow

The agent loads shared context.

Use the approved coding route,
load the backend release skill,
then attach spend evidence.

→ route selected
→ skill packet loaded
→ policy checked
→ report recorded
Outcome stays inside the agent workflow.

Workflow context

Skills travel with the work.

Attach release rules, customer policy, budget limits, and reporting labels to repeated workflows so every agent starts with the approved packet.

InputShared skillRoute choiceSpend report

Shared library

Team knowledge, one call away.

Approved model groups, pricing estimates, shared skills, and reporting tools can be exposed as one compact Quotaflow catalog.

GPT routesopenai:gpt-5.5Claude routesanthropic:claude-sonnetGemini routesgoogle:gemini-flashOpen source safequotaflow:oss-safeImage routesopenai:gpt-image-2Video routesgoogle:veo / seedanceEmbeddingstext-embedding-3Spend reportquotaflow:usageShared skillsquotaflow:skillsPolicy mapquotaflow:policySee all model routes ↗

Team context

Where shared context earns its place.

Use Quotaflow where agents already work: coding, automations, route previews, and repeated operations.

Inside your coding agent

Ask the agent to choose the route and attach the context.

Example: review this ticket using the backend release skill, choose a policy-approved lower-cost route where safe, and report estimated savings.

decision: claude-sonnet route\nskill: backend-release-checklist\nreport: attached to workflow\nreport: workspace / engineering
Routeapproved
Skillloaded
Reportattached
Ownervisible

Production

Built to run on your account.

Your keys stay server-side.

Provider credentials stay behind the gateway, not inside every agent client.

Discounts are scoped.

Pricing depends on package, route, and terms. Confirm larger discounts with sales.

Same tools as production.

Shared context packets, skills, shared context, and spend reports match the production control plane.

FAQ

It is the reusable knowledge and context layer inside the Quotaflow wallet. Agents can load approved skills, project context, route policy, budget labels, and usage reports instead of rebuilding prompts in every thread.

Quotaflow Quotaflow

Give every agent the same approved context.

Start with one repeated workflow. Confirm shared skills, route policy, and reporting before rolling out wider.