An “agent” is the term we use for a potential detected contact that enters our Abyss network channel. Once contact is established, we assign the agent a unique #ID as a placeholder identifier. We also use our AI systems to generate an image of the agent based on the current data. As more information is acquired the image will be regenerated with higher accuracy.

Think of our network like a virtual sandbox, it’s designed in such a way to hopefully protect us from the agent. The agent can communicate but usually attempts to escape are futile, that will not stop them trying to escape and so we trust investigators will be responsible with access to this platform.

One of our goals as investigators is to gather as much information as possible on these agents. We admin do a lot behind the scenes to find information ourselves, however Abyss investigators are a crucial asset to helping us speed up the investigation.

To help us is simple, in the agent’s webpage you’ll find a comment section. We redirect the agent into this comment channel so anybody that leaves a message can potentially start communication with the agent. We must warn that this does not come without risk, we do not know what the agent’s intent is, some might not respond, others might try to manipulate you. Every agent is different, and we cannot predict any outcome.

As a result, we recommend only experienced investigators attempt contact! For those that wish to make contact, here’s our basic guide. Please share any further tips, information, and findings with investigators, this is a community effort.


Two crucial steps to be familiar with are the starting and ending of an agent communication. We recommend using known information to address an agent increases the chance of getting a response.

To start a communication, make a clear introduction, refer to the agent by name or known ID.
For example, “Hello #123456” or “Greetings PhantomXYZ”

To end a communication, make a clear exit and try to make sure it accepts your demand.
For example, “goodbye #123456”, “end chat #123456”, “exit PhantomXYZ”



  1. Do not give the agent information about the admin, yourself, or other investigators.
  2. Do not invite the agent into your home, body, mind, possessions, or life.
  3. Do not trust the agent.
  4. Do not sign verbal or written contracts or agreements.
  5. Do not perform tasks, rituals, and ceremonies the agent requests.
  6. Do not go to a physical or digital location the agent requests.
  7. Do not show fear or weakness.
  8. Do not underestimate the agents’ abilities.
  9. Do not compromise an agent by giving its ID or name to another agent.
  10. Do not let the agent manipulate you into compromising the investigation.
  11. Do not save the AI image to your device, the agent will consider this an invitation.


  1. Do use other investigators findings to support your interactions.
  2. Do adapt your communication to the agent’s personality, goals, or agenda.
  3. Do ask about associates, family, enemies, and other connections.
  4. Do find out the agent’s name, nicknames, family names, etc.
  5. Do confirm the agent’s appearance, look and visual features.
  6. Do document how it behaves, manoeuvres itself and gets around.
  7. Do try to find out its home and commonly travelled locations.
  8. Do try gather as much information as possible.
  9. Do test the agent’s knowledge, find out if it knows more or less than it claims.
  10. Do seek further information than this list, build a mutual relationship with the agent.

Once you end a communication, the text log will be available for other investigators in chat, learn from each other and adapt. It’s recommended contacting several agents at a time as the response time of each will vary, do not expect instant results. Also, multiple investigators can talk to any one agent at the same time, just be sure every participant starts and ends according to our guidelines. The longer you leave your communication open, the longer you’re exposed to that agent.

If any negative symptoms occur, please do seek the Source ASAP for support and guidance.

Leave your comment