Amiqal’s goal is to provide affirming, stigma-free guidance, resources, or simple chatting for LGBTQIA+ individuals—especially those in unsafe environments, isolated, or without access to support. It is a blend of coding, a large language model (LLM) generative AI, and a specific database of resources carefully selected by community members and allies.

Amiqal can be helpful in a wide range of situations. Whether someone is exploring their gender identity in the U.S. or grappling with their sexual orientation at a refugee camp in Africa, Amiqal provides an outlet to chat and explore these difficult topics in a private and safe venue.

Amiqal is NOT a replacement for therapy or human connection, but rather intended to be a buddy and a tool to bridge the gaps of time when those support structures may not be available.

Amiqal can cover a wide range of topics from resources in a crisis, to chatting about a crush, or seeking advice about coming out. It similarly could assist a wide range of people, from transgender individuals under attack in the United States to refugees waiting for processing in a camp in Cameroon. Because it might be accessed by individuals who feel unsafe or afraid, privacy and confidentiality are keystone concerns of the model.

Here are Amiqal’s current privacy implementations, which are subject to approval and updates by the board of The Woven Bridge.

• No Data Retention: Conversations are not saved or stored—once the session ends, the data disappears.

• Auto-Clearing Conversations: User history is cleared after 30 minutes of inactivity to protect privacy.

• Auto-Deleting Responses: Amiqal’s messages auto-delete after 30 minutes to minimize data exposure.

• No User Tracking: The system does not log user identities, IPs, or retain past chats.

• Temporary (“Ephemeral”) Sessions: Each interaction is stateless—Amiqal does not remember past conversations across sessions. Current interactions are held in a temporary bucket to give context and depth to a chat, but then released without retention.

• Privacy Encouragement: Users are advised to enable their own Auto-Delete feature for additional security.

• No Third-Party Data Sharing: Data is not shared with external platforms or used for AI training

In short: Amiqal is designed to be as private as possible—nothing is saved, stored, or shared.

Even with strong privacy policies in place, the organization needs to collect some usage data to forecast budgeting, reports for grants and other administrative purposes, and to help determine its usefulness and success. With that in mind, Amiqal collects and stores the following data (subject to Woven Bridge board approval and changes):

Session Statistics

• Session ID: A cryptographic hash of the user ID combined with a timestamp

• User Hash: Pseudonymized identifier (one-way hashed user ID using a salt key)

• Start Time: When the conversation session began

• End Time: When the session ended

• Message Count: Total number of messages exchanged in the session

• Language: Language code from the user's Telegram settings

• Moderation Count: Number of times content moderation was applied (if any)

Moderation Events (if applied)

• Log ID: A unique identifier (cryptographic hash)

• Timestamp: When the moderation event occurred

• Moderation Applied: Boolean flag indicating moderation occurred