Agents FAQ
Answers to frequently asked questions about Agents.
| Agents Availability | |||
|---|---|---|---|
| Cloud | Public beta | ||
| Self-hosted Edge 3.234 or later | Public beta | ||
| Self-hosted Stable 3.253 or later | Public beta | ||
This FAQ aims to provide answers to common questions about Agents, and provide you with the information you need to leverage Agents effectively.
General
Frequently asked general questions about Agents:
Where can I provide feedback about Agents?
Please share feedback in Retool’s community forum.
How can I get support for Agents?
Reach out to your Enterprise account manager or contact Retool support.
Where can I go to learn more about Agents?
The following links provide more information about AI and Agents.
Data collection and privacy
Frequently asked questions about data collection and how information is used with Agents:
Is my data ever used to train or fine-tune models?
No, Retool uses foundation models. Neither Retool nor the providers of the third-party models will use agent inputs or outputs to train or fine-tune models. Prior to making a third-party model available for use with Agents, Retool ensures contractual safeguards are in place that restrict such providers from using inputs and outputs for their own purposes.
What about Retool’s other AI vendors?
Retool engages other vendors to help provide agents. Vendors that process agent inputs and outputs on Retool’s behalf are Retool’s sub-processors and are also contractually restricted from using inputs and outputs for their own purposes.
If I use an open-source model, will the model provider have access to my inputs or outputs?
No, Retool-managed, open-source models are hosted by Retool’s third-party infrastructure provider. Inputs and outputs will not be shared with the provider of the open-source model.
What data is sent to the Retool-managed agents observability API?
The data sent to the Retool-managed (cloud-hosted) agents observability API consists of the following:
- UUIDs corresponding to objects in the Retool deployment's Postgres database.
- Success/failure status of agent runs.
- Model type of agent runs.
- Runtime of agent runs.
- Number of LLM tokens used by agent runs.
- Information about the tools used during agent runs. Tool information is limited to UUIDs and the configured tool name and tool parameter names.
No customer data beyond the names of configured tools and tool parameters are sent to this API.
If you have any feedback or concerns about this data, please reach out to your account team.
AI models and LLM keys
Frequently asked questions about AI models and LLM-provider keys:
Which providers and models are available?
Retool-managed keys are available for certain AI providers and models. If you use Agents with any of these third-party models, the third-party model will process all agent inputs, including the prompt provided by the internal user of the agent, to initiate performance of a task or to generate output. Please review and comply with any policies published by the provider of the third-party model.
Can I bring my existing LLM-provider keys?
Yes. When you bring your own key to an account with a model provider, the provider’s handling of data while under its control will be governed by the agreement you have with that provider. When output data that is returned by the provider is in Retool’s control, such data will be subject to your agreement with Retool.
Self-hosted
Frequently asked questions about Agents when you self-host Retool:
What are the differences between the edge release and a standard deployment of Retool?
The infrastructure of the Agents edge deployment is very similar to a Retool deployment with workflows. The following example code block shows the agent-worker and agent-eval-worker container configuration for a Docker Compose deployment.
services:
# Temporal workers for Agents
agent-worker:
build:
context: .
env_file: docker.env
environment:
- SERVICE_TYPE=WORKFLOW_TEMPORAL_WORKER
networks:
- backend
- code-executor
depends_on:
- postgres
restart: always
agent-eval-worker:
build:
context: .
env_file: docker.env
environment:
- SERVICE_TYPE=AGENT_EVAL_TEMPORAL_WORKER
networks:
- backend
- code-executor
depends_on:
- postgres
restart: always
Using a Retool-managed agents observability API is required in the edge release. This API strictly does not receive, persist, or have access to sensitive customer data, agent inputs, or agent outputs. It is required to power the Monitor dashboard.
How do I set up evals for self-hosted instances?
Evals work in self-hosted deployments without any special configuration. To optimize performance, you can configure Amazon Simple Storage Service (Amazon S3) to store test case inputs and outputs (which can be large) instead of Postgres. To use Amazon S3, first configure Amazon S3 as a resource, and then set the following environment variables:
# Replace the following example values with your S3 values.
AGENT_EVALS_S3_BUCKET='retool-agent-evals'
AGENT_EVALS_S3_ACCESS_KEY_ID='AKIAIOSFODNN7EXAMPLE'
AGENT_EVALS_S3_SECRET_ACCESS_KEY='wJalrXUtnFEMI/K7MDENG/bPxRfiCYEXAMPLEKEY'
AGENT_EVALS_S3_REGION='us-west-2'
Does the edge release send any agent inputs and outputs to Retool servers?
No agent inputs or outputs are sent back to Retool servers.
In which releases are Agents included?
Agents is in public beta with the 3.234.0 edge release and the 3.253.0 stable release for Self-hosted organizations.