Survey | Trustabl.ai

Tell us about your setup

1. What framework(s) are you using to build your agents? pick all that apply

1. What framework(s) are you using to build your agents?

Other

2. What language(s) are your agents written in? pick all that apply

2. What language(s) are your agents written in?

Python

TypeScript / JavaScript

Java / Kotlin

Other

3. What tools or integrations do your agents use?

e.g. internal APIs, databases, Slack, payment/refunds, search, email…

3. What tools or integrations do your agents use?

Do you connect any of these via MCP (Model Context Protocol)?

Yes

Not Sure

4. Where do your agents actually run? pick all that apply

4. What language(s) are your agents written in?

Locally / on a developer machine

A backend server we manage (VM, bare metal)

Serverless (AWS Lambda, Google Cloud Run, Azure Functions)

Containers (ECS, Kubernetes, Fargate)

A PaaS (Vercel, Render, Railway, Fly.io)

A managed agent platform

Other

5. How do you ship changes to your agents? pick all that apply

5. What language(s) are your agents written in?

GitHub Actions

GitLab CI

Jenkins

Bitbucket Pipelines

Manual deploy / no CI

Other

6. What stage are your agents at? pick all that apply

6. What language(s) are your agents written in?

Just experimenting / prototyping

Testing internally

In production

What we already do

Quick context before you choose. Trustabl already scans your tools and generates the basics automatically:

✓

Production readiness score

For each tool

✓

Input validation & clean schemas

MCP / OpenAI / Claude compatible

✓

Retry & idempotency safety

Built-in resilience checks

✓

Core documentation & examples

Auto-generated from your tools

Pick your top 5

Out of everything below, which 5 would help you the most right now? Don’t overthink it — pick the ones that match your real pain.

0 / 5 selected

Smarter tool selection

Retry safety— Flags whether a tool is safe to retry so your agent won't double-charge or duplicate actions

Error catalog— Common errors + how to fix them, so the agent recovers instead of looping

Timeouts & fallbacks— What the agent should do when a tool hangs or fails

Reliability — stop agents from breaking

When to use / when NOT to use— Clear rules so the agent picks the right tool

Tool purpose— A plain one-liner on what each tool is actually for

Prompt snippets— Copy-paste tool descriptions for your system prompt

Input & output safety

Input validation rules— Blocks malformed or invalid tool calls before they run

Output schema— Predictable, structured responses your agent can rely on

Visibility — know what your agent is doing

Tracing (OTEL)— See every step the agent took

Structured logging— Clean logs with automatic PII redaction

Usage metrics— Counts, latencies, and failure rates per tool

Safety, policy & compliance

Approval gates— Require human sign-off for risky actions (e.g. refunds over $5K)

Risk level— Flags how risky each tool is to call

Data / PII handling— How sensitive data is classified and protected

Audit logs— Compliance-ready logging on every call

Trust & provenance

Signing / attestation— Tamper-proof proof of what a tool is and does (Sigstore)

OpenShell sandbox policies— Auto-generate least-privilege runtime policies

Compatibility & exports

Export formats— Ready-to-use MCP / OpenAI / Claude / LangChain / GitAgent definitions

Cost & performance

Cost per call estimate— See roughly what each tool costs (USD / tokens)

Latency profile— Expected speed of each tool

Caching recommendations— Cut repeat calls and save tokens

Something we're missing?

One open question

What's the #1 thing that breaks your agents in production?

Want early access?

Want first access to the fields you picked once we ship them? Drop your email (optional) and we'll let you know — and tell you when your requests go live.

Help shape what Trustabl builds next

Tell us about your setup

1. What framework(s) are you using to build your agents? pick all that apply

2. What language(s) are your agents written in? pick all that apply

3. What tools or integrations do your agents use?

4. Where do your agents actually run? pick all that apply

5. How do you ship changes to your agents? pick all that apply

6. What stage are your agents at? pick all that apply

What we already do

Pick your top 5

One open question

Want early access?