SEARCH TOOL CATALOGRUNNING
MOUNT MCP CONNECTIONSRUNNING
EXECUTE CONNECTED TOOLRUNNING
Everything your agents need to take action.
1,400 apps. 15,000 tools. One secure connection layer for every agent.
01READING PROMPTClaude Code
Now showing Claude Code
■ WHY VIA
Give agents a gateway
they can rely on.
⌕VIA tool catalog
›list sentry errors and create linear issues
resolving intent...
SENTRY
List unresolved issues
LINEAR
Create issue
2 Actions selected · execution plan ready
01
A catalog built for execution
Agents discover the right provider and Action, inspect its contract, then execute through one consistent gateway.
- 1,400 apps and 15,000 prebuilt tools
- Request, response, scope, and credential metadata
- Stable tool IDs across every workspace
Contract coverage
3 Actions
schema ready
SENTRY_LIST_ISSUES99.4%
LINEAR_CREATE_ISSUE98.8%
GITHUB_GET_COMMIT97.9%
1.4Kapps
15Ktools
1layer
02
One contract, every client
Expose the same Actions to MCP hosts, application code, local agents, and custom API clients without rewriting integrations.
- MCP, Connector SDK, oo CLI, HTTP, and OpenAPI
- Lazy-loaded provider executors
- Portable contracts with no framework lock-in
Agent chatconnected
Review Sentry incidents and create Linear issues for anything critical.
Authorization required
Connect Linear to continue. Access is scoped to creating issues only.
Ask your agent something...↑
03
Connect accounts once
Keep provider secrets inside the runtime while agents receive safe metadata and the exact scoped access required for each run.
- OAuth2, API keys, custom credentials, and no-auth providers
- Connection aliases, scopes, runtime tokens, and policies
- Encrypted credentials never enter the agent process
Inspect & run Actionsruntime · policy enforced
issues = run_connector_action(
'SENTRY_LIST_ISSUES',
project='api-prod', status='unresolved'
)
for issue in issues:
issue['trace'] = run_connector_action(
'SENTRY_GET_EVENT', issue_id=issue['id']
)
ranked = invoke_llm('Classify P0/P1/P2')runtime ready · logs redacted
04
Own the runtime
Run your connection layer locally, in your cloud, or through a managed workspace without changing the tool contract.
- Docker, Node.js, Fly.io, Kubernetes, and Cloudflare Workers
- SQLite, PostgreSQL, D1, R2, or S3-compatible storage
- Redacted run logs, health checks, and inspectable state
CONNECT ONCE TO FULL CONTROL
One runtime, every workflow
VIA
FOR YOUConnect any agent host or application to the services your users already rely on. One gateway turns requests into stable, inspectable Actions.
Credentials stay behind the runtime while scopes, policies, and run logs remain fully inspectable.
codex — ~/Desktop/jcode
>_OpenAI Codexv0.149.1
model:gpt-5.6-sol high /model to changedirectory:~/Desktop/jcode
•Connected to VIA tools
•Indexed repository context
•Ready to edit, run, and verify
›Ask Codex to do anything
? for shortcutsgpt-5.6-sol high · ~/Desktop/jcode
ChatWork
✦What should we work on?
Work on anything
+
5.6 Sol Light⌄♩↑
▢ Project ▥ Files M Plugins▣ Get desktop app
✦OpenAI Work
WorkspaceOpenAI Work
Work with your team’s context
Files · sources · connected tools
Prepare a polished deliverable
›Create a launch plan from these files
Ready to beginWorkspace Connected
Runtime APIs
PLATFORMGive every agent secure access to 1,400 apps and 15,000 tools through one consistent runtime.
AI & agents
142Models, agents, search, and media generation
Commerce & payments
124Storefronts, orders, billing, and fulfillment
Team collaboration
236Messaging, email, meetings, and notifications
Docs & knowledge
154Documents, knowledge bases, drives, and spreadsheets
Projects & productivity
247Projects, tasks, schedules, and automation
Marketing & growth
162Ads, content, campaigns, and growth data
Data & analytics
171Warehouses, databases, BI, and storage
Development & cloud
164Code, deployment, cloud services, and observability
+ For builders
VIA
Explore the platformCredential Boundary
Keep OAuth tokens, API keys, scopes, and connection identities behind one inspectable runtime.
Learn more
Every Agent Interface
Expose the same Actions through MCP, TypeScript, CLI, HTTP, and generated OpenAPI.
Learn more
Stable Action Contracts
result = connector.actions.execute((
"github.get_current_user",
{ connection: "work" }
)
Inspect schemas, required scopes, policies, and execution results before an Action reaches production.
Portable by Design
Run locally, on Fly.io, Kubernetes, Cloudflare, or a managed workspace while keeping identical provider and Action IDs.
+ Runtime controls
Keep credentials private. Keep every run inspectable.
Scoped runtime tokens+
Keep provider secrets outside the agent process with scoped connections, policy enforcement, and auditable execution.
Action allow and block policies+
Keep provider secrets outside the agent process with scoped connections, policy enforcement, and auditable execution.
Private deployment+
Keep provider secrets outside the agent process with scoped connections, policy enforcement, and auditable execution.
Redacted, inspectable run logs+
Keep provider secrets outside the agent process with scoped connections, policy enforcement, and auditable execution.