PlayerEngine + Player2NPC — v1.3.0 Release Notes

Minecraft 1.20.1 (Fabric · Forge) | Minecraft 1.21.1 (Fabric · NeoForge)

---

🧠 Your Companion Can Now Plan and Execute Multi-Step Goals

The biggest addition in 1.3.0 is the agentic planning system. Instead of issuing one command at a time, you can now give your companion a natural-language goal and it will figure out the steps on its own.

Tell your companion something like:

"Pick up all the slimeballs on the ground and store them in a chest"

It will gather the drops, find a suitable chest, and deposit everything — on its own, without you managing each step. If something goes wrong mid-plan, it tells you what failed rather than silently stopping.

The planner uses a lightweight built-in fallback for common patterns (gather → store, find a chest → deposit) so simple goals don't always require an AI call. More complex or open-ended goals go through the AI planner.

---

🔨 Crafting From Scratch

Companions can now craft items for you. Give them a crafting goal and they will figure out what materials are needed, go gather anything that's missing, find or place a crafting table, and work through whatever steps the recipe requires — including multi-step chains like turning logs into planks into sticks.

This works naturally alongside the planning system, so a plan can include a crafting step mixed in with gathering and storing.

---

📦 Storage Memory — Your Companion Remembers Where Things Are

Companions now build an internal map of storage containers they interact with, powered by a new system called EllieGPS. Over time your companion learns which chests hold which categories of items and can go directly to the right one when a plan calls for it.

Tell your companion to scan a nearby chest and it will learn its contents. Ask it to put something away and it will find the right chest for it. Ask it where you keep your iron and it will tell you or go get it. The memory persists across play sessions, so your base layout is remembered even after relogging.

Waypoint management is also available — you can ask your companion to mark a chest location, look up what it knows about your storage, or clear outdated entries.

---

💰 Budget Controls

New /player2npc budget commands let you control how much your companions are allowed to spend on AI within a rolling time window. Set a limit, and the mod enforces it automatically — no surprises on your Joules balance after a long session.

/player2npc budget status — shows your current limits, how many AI calls have been made this window, and your cached Joules balance.

/player2npc budget reset — clears the current call window immediately so AI requests are allowed again. Useful if you hit a limit and want to keep going.

/player2npc budget window <minutes> — sets how long the rolling window lasts before call counts reset. For example, window 60 gives you a fresh allowance every hour.

/player2npc budget soft <calls> — sets a soft call limit. When hit, the mod may route to a lighter AI model instead of stopping entirely. Set to 0 to disable.

/player2npc budget hard <calls> — sets a hard call limit. When hit, AI calls stop completely until the window resets. Set to 0 to disable.

/player2npc budget joules_soft <amount> — same idea as the soft call limit, but based on your Joules balance rather than call count. Triggers when your balance drops to or below this amount.

/player2npc budget joules_hard <amount> — hard stop based on Joules balance. AI calls are blocked until your balance recovers above this threshold.

/player2npc budget joules_refresh <seconds> — controls how often your Joules balance is re-checked from the Player2 platform. Defaults are conservative to avoid unnecessary API calls.

On servers where the owner pays for AI costs rather than individual players, operators can manage limits from the engine side with /playerengine player2 budget.

---

🤖 AI & Power User Updates

Smarter Tool Retrieval

The AI now uses on-device semantic search to pick the right tool for a given request, rather than loading the full command list into every prompt. This keeps things efficient as the available command set grows, and generally makes tool selection more accurate.

Learning From Mistakes

When the AI retrieves the wrong tasks, it automatically retries with a rephrased query and records the failure as a learning signal. Over time, the system builds up accurate aliases for your usage patterns so the same retrieval miss doesn't repeat. Currently off by default. Turn on in /playerengine/server_player2.json (Deepsearch and related settings)

Model Tier Routing

AI calls are now automatically routed to the appropriate model tier based on the task. Heavy planning calls can use a stronger model; lightweight follow-ups use a cheaper one. This happens automatically based on your profile and patron tier — nothing to configure.

Mod Intelligence

The companion now scans the mods installed on your server and builds a background knowledge base of their items, blocks, and entities. This runs quietly after world load and improves over time. Companions with this knowledge can reason about modded content more accurately — recognizing what an unfamiliar modded item does rather than treating it as completely unknown.

You can check the status or trigger a rebuild with /playerengine capability status. See the enrichment guide for setup and tuning details.

Budget Guards

The budget enforcement runs at the engine level on every AI call — in singleplayer and on servers alike. Patrons with a named profile can designate it as a fallback that the engine automatically switches to when a soft limit is hit, so usage can step down to a cheaper model before stopping entirely. On dedicated servers where each player pays for their own usage, limits track per-player rather than globally.

---

🐛 Bug Fixes

Crash on World Load With Certain Mods

An internal hook was triggering crashes when loading into worlds that contained mods with complex item behavior — notably SilentGear and Construction Sticks in modpacks like All the Mods 10. The game would crash before gameplay even started. Fixed.

Multiple Companions Talking Over Each Other

When more than one companion was active, they would respond simultaneously — a companion would start speaking before the previous one had finished. Companions now wait for the current speaker's audio to complete before queuing their own response. Fixed.

Companion Wandering Instead of Moving to a Target

An overflow bug in the companion's movement system caused it to treat any block approach as already complete, so it would just wander instead of walking up to what it needed to reach. This affected any task involving moving to a specific block — chests, crafting tables, resource nodes. Fixed.

---

⚠️ Known Issues

gamer command (autonomous "beat the game" run) is temporarily disabled. It will return in a future update once outstanding issues with its execution are resolved.

Voice transcription (STT) may produce empty results for some users. If your companion isn't responding to voice input but responds fine to typed chat, check latest.log for [STT] entries — this will help us narrow down the cause.

PlayerEngine + Player2NPC — v1.3.0 Release Notes

Minecraft 1.20.1 (Fabric · Forge) | Minecraft 1.21.1 (Fabric · NeoForge)

---

🧠 Your Companion Can Now Plan and Execute Multi-Step Goals

The biggest addition in 1.3.0 is the agentic planning system. Instead of issuing one command at a time, you can now give your companion a natural-language goal and it will figure out the steps on its own.

Tell your companion something like:

"Pick up all the slimeballs on the ground and store them in a chest"

It will gather the drops, find a suitable chest, and deposit everything — on its own, without you managing each step. If something goes wrong mid-plan, it tells you what failed rather than silently stopping.

The planner uses a lightweight built-in fallback for common patterns (gather → store, find a chest → deposit) so simple goals don't always require an AI call. More complex or open-ended goals go through the AI planner.

---

🔨 Crafting From Scratch

Companions can now craft items for you. Give them a crafting goal and they will figure out what materials are needed, go gather anything that's missing, find or place a crafting table, and work through whatever steps the recipe requires — including multi-step chains like turning logs into planks into sticks.

This works naturally alongside the planning system, so a plan can include a crafting step mixed in with gathering and storing.

---

📦 Storage Memory — Your Companion Remembers Where Things Are

Companions now build an internal map of storage containers they interact with, powered by a new system called EllieGPS. Over time your companion learns which chests hold which categories of items and can go directly to the right one when a plan calls for it.

Tell your companion to scan a nearby chest and it will learn its contents. Ask it to put something away and it will find the right chest for it. Ask it where you keep your iron and it will tell you or go get it. The memory persists across play sessions, so your base layout is remembered even after relogging.

Waypoint management is also available — you can ask your companion to mark a chest location, look up what it knows about your storage, or clear outdated entries.

---

💰 Budget Controls

New /player2npc budget commands let you control how much your companions are allowed to spend on AI within a rolling time window. Set a limit, and the mod enforces it automatically — no surprises on your Joules balance after a long session.

/player2npc budget status — shows your current limits, how many AI calls have been made this window, and your cached Joules balance.

/player2npc budget reset — clears the current call window immediately so AI requests are allowed again. Useful if you hit a limit and want to keep going.

/player2npc budget window <minutes> — sets how long the rolling window lasts before call counts reset. For example, window 60 gives you a fresh allowance every hour.

/player2npc budget soft <calls> — sets a soft call limit. When hit, the mod may route to a lighter AI model instead of stopping entirely. Set to 0 to disable.

/player2npc budget hard <calls> — sets a hard call limit. When hit, AI calls stop completely until the window resets. Set to 0 to disable.

/player2npc budget joules_soft <amount> — same idea as the soft call limit, but based on your Joules balance rather than call count. Triggers when your balance drops to or below this amount.

/player2npc budget joules_hard <amount> — hard stop based on Joules balance. AI calls are blocked until your balance recovers above this threshold.

/player2npc budget joules_refresh <seconds> — controls how often your Joules balance is re-checked from the Player2 platform. Defaults are conservative to avoid unnecessary API calls.

On servers where the owner pays for AI costs rather than individual players, operators can manage limits from the engine side with /playerengine player2 budget.

---

🤖 AI & Power User Updates

Smarter Tool Retrieval

The AI now uses on-device semantic search to pick the right tool for a given request, rather than loading the full command list into every prompt. This keeps things efficient as the available command set grows, and generally makes tool selection more accurate.

Learning From Mistakes

When the AI retrieves the wrong tasks, it automatically retries with a rephrased query and records the failure as a learning signal. Over time, the system builds up accurate aliases for your usage patterns so the same retrieval miss doesn't repeat. Currently off by default. Turn on in /playerengine/server_player2.json (Deepsearch and related settings)

Model Tier Routing

AI calls are now automatically routed to the appropriate model tier based on the task. Heavy planning calls can use a stronger model; lightweight follow-ups use a cheaper one. This happens automatically based on your profile and patron tier — nothing to configure.

Mod Intelligence

The companion now scans the mods installed on your server and builds a background knowledge base of their items, blocks, and entities. This runs quietly after world load and improves over time. Companions with this knowledge can reason about modded content more accurately — recognizing what an unfamiliar modded item does rather than treating it as completely unknown.

You can check the status or trigger a rebuild with /playerengine capability status. See the enrichment guide for setup and tuning details.

Budget Guards

The budget enforcement runs at the engine level on every AI call — in singleplayer and on servers alike. Patrons with a named profile can designate it as a fallback that the engine automatically switches to when a soft limit is hit, so usage can step down to a cheaper model before stopping entirely. On dedicated servers where each player pays for their own usage, limits track per-player rather than globally.

---

🐛 Bug Fixes

Crash on World Load With Certain Mods

An internal hook was triggering crashes when loading into worlds that contained mods with complex item behavior — notably SilentGear and Construction Sticks in modpacks like All the Mods 10. The game would crash before gameplay even started. Fixed.

Multiple Companions Talking Over Each Other

When more than one companion was active, they would respond simultaneously — a companion would start speaking before the previous one had finished. Companions now wait for the current speaker's audio to complete before queuing their own response. Fixed.

Companion Wandering Instead of Moving to a Target

An overflow bug in the companion's movement system caused it to treat any block approach as already complete, so it would just wander instead of walking up to what it needed to reach. This affected any task involving moving to a specific block — chests, crafting tables, resource nodes. Fixed.

---

⚠️ Known Issues

gamer command (autonomous "beat the game" run) is temporarily disabled. It will return in a future update once outstanding issues with its execution are resolved.

Voice transcription (STT) may produce empty results for some users. If your companion isn't responding to voice input but responds fine to typed chat, check latest.log for [STT] entries — this will help us narrow down the cause.

PlayerEngine + Player2NPC — v1.3.0 Release Notes

Minecraft 1.20.1 (Fabric · Forge) | Minecraft 1.21.1 (Fabric · NeoForge)

---

🧠 Your Companion Can Now Plan and Execute Multi-Step Goals

The biggest addition in 1.3.0 is the agentic planning system. Instead of issuing one command at a time, you can now give your companion a natural-language goal and it will figure out the steps on its own.

Tell your companion something like:

"Pick up all the slimeballs on the ground and store them in a chest"

It will gather the drops, find a suitable chest, and deposit everything — on its own, without you managing each step. If something goes wrong mid-plan, it tells you what failed rather than silently stopping.

The planner uses a lightweight built-in fallback for common patterns (gather → store, find a chest → deposit) so simple goals don't always require an AI call. More complex or open-ended goals go through the AI planner.

---

🔨 Crafting From Scratch

Companions can now craft items for you. Give them a crafting goal and they will figure out what materials are needed, go gather anything that's missing, find or place a crafting table, and work through whatever steps the recipe requires — including multi-step chains like turning logs into planks into sticks.

This works naturally alongside the planning system, so a plan can include a crafting step mixed in with gathering and storing.

---

📦 Storage Memory — Your Companion Remembers Where Things Are

Companions now build an internal map of storage containers they interact with, powered by a new system called EllieGPS. Over time your companion learns which chests hold which categories of items and can go directly to the right one when a plan calls for it.

Tell your companion to scan a nearby chest and it will learn its contents. Ask it to put something away and it will find the right chest for it. Ask it where you keep your iron and it will tell you or go get it. The memory persists across play sessions, so your base layout is remembered even after relogging.

Waypoint management is also available — you can ask your companion to mark a chest location, look up what it knows about your storage, or clear outdated entries.

---

💰 Budget Controls

New /player2npc budget commands let you control how much your companions are allowed to spend on AI within a rolling time window. Set a limit, and the mod enforces it automatically — no surprises on your Joules balance after a long session.

/player2npc budget status — shows your current limits, how many AI calls have been made this window, and your cached Joules balance.

/player2npc budget reset — clears the current call window immediately so AI requests are allowed again. Useful if you hit a limit and want to keep going.

/player2npc budget window <minutes> — sets how long the rolling window lasts before call counts reset. For example, window 60 gives you a fresh allowance every hour.

/player2npc budget soft <calls> — sets a soft call limit. When hit, the mod may route to a lighter AI model instead of stopping entirely. Set to 0 to disable.

/player2npc budget hard <calls> — sets a hard call limit. When hit, AI calls stop completely until the window resets. Set to 0 to disable.

/player2npc budget joules_soft <amount> — same idea as the soft call limit, but based on your Joules balance rather than call count. Triggers when your balance drops to or below this amount.

/player2npc budget joules_hard <amount> — hard stop based on Joules balance. AI calls are blocked until your balance recovers above this threshold.

/player2npc budget joules_refresh <seconds> — controls how often your Joules balance is re-checked from the Player2 platform. Defaults are conservative to avoid unnecessary API calls.

On servers where the owner pays for AI costs rather than individual players, operators can manage limits from the engine side with /playerengine player2 budget.

---

🤖 AI & Power User Updates

Smarter Tool Retrieval

The AI now uses on-device semantic search to pick the right tool for a given request, rather than loading the full command list into every prompt. This keeps things efficient as the available command set grows, and generally makes tool selection more accurate.

Learning From Mistakes

When the AI retrieves the wrong tasks, it automatically retries with a rephrased query and records the failure as a learning signal. Over time, the system builds up accurate aliases for your usage patterns so the same retrieval miss doesn't repeat. Currently off by default. Turn on in /playerengine/server_player2.json (Deepsearch and related settings)

Model Tier Routing

AI calls are now automatically routed to the appropriate model tier based on the task. Heavy planning calls can use a stronger model; lightweight follow-ups use a cheaper one. This happens automatically based on your profile and patron tier — nothing to configure.

Mod Intelligence

The companion now scans the mods installed on your server and builds a background knowledge base of their items, blocks, and entities. This runs quietly after world load and improves over time. Companions with this knowledge can reason about modded content more accurately — recognizing what an unfamiliar modded item does rather than treating it as completely unknown.

You can check the status or trigger a rebuild with /playerengine capability status. See the enrichment guide for setup and tuning details.

Budget Guards

The budget enforcement runs at the engine level on every AI call — in singleplayer and on servers alike. Patrons with a named profile can designate it as a fallback that the engine automatically switches to when a soft limit is hit, so usage can step down to a cheaper model before stopping entirely. On dedicated servers where each player pays for their own usage, limits track per-player rather than globally.

---

🐛 Bug Fixes

Crash on World Load With Certain Mods

An internal hook was triggering crashes when loading into worlds that contained mods with complex item behavior — notably SilentGear and Construction Sticks in modpacks like All the Mods 10. The game would crash before gameplay even started. Fixed.

Multiple Companions Talking Over Each Other

When more than one companion was active, they would respond simultaneously — a companion would start speaking before the previous one had finished. Companions now wait for the current speaker's audio to complete before queuing their own response. Fixed.

Companion Wandering Instead of Moving to a Target

An overflow bug in the companion's movement system caused it to treat any block approach as already complete, so it would just wander instead of walking up to what it needed to reach. This affected any task involving moving to a specific block — chests, crafting tables, resource nodes. Fixed.

---

⚠️ Known Issues

gamer command (autonomous "beat the game" run) is temporarily disabled. It will return in a future update once outstanding issues with its execution are resolved.

Voice transcription (STT) may produce empty results for some users. If your companion isn't responding to voice input but responds fine to typed chat, check latest.log for [STT] entries — this will help us narrow down the cause.

PlayerEngine + Player2NPC — v1.3.0 Release Notes

Minecraft 1.20.1 (Fabric · Forge) | Minecraft 1.21.1 (Fabric · NeoForge)

---

🧠 Your Companion Can Now Plan and Execute Multi-Step Goals

The biggest addition in 1.3.0 is the agentic planning system. Instead of issuing one command at a time, you can now give your companion a natural-language goal and it will figure out the steps on its own.

Tell your companion something like:

"Pick up all the slimeballs on the ground and store them in a chest"

It will gather the drops, find a suitable chest, and deposit everything — on its own, without you managing each step. If something goes wrong mid-plan, it tells you what failed rather than silently stopping.

The planner uses a lightweight built-in fallback for common patterns (gather → store, find a chest → deposit) so simple goals don't always require an AI call. More complex or open-ended goals go through the AI planner.

---

🔨 Crafting From Scratch

Companions can now craft items for you. Give them a crafting goal and they will figure out what materials are needed, go gather anything that's missing, find or place a crafting table, and work through whatever steps the recipe requires — including multi-step chains like turning logs into planks into sticks.

This works naturally alongside the planning system, so a plan can include a crafting step mixed in with gathering and storing.

---

📦 Storage Memory — Your Companion Remembers Where Things Are

Companions now build an internal map of storage containers they interact with, powered by a new system called EllieGPS. Over time your companion learns which chests hold which categories of items and can go directly to the right one when a plan calls for it.

Tell your companion to scan a nearby chest and it will learn its contents. Ask it to put something away and it will find the right chest for it. Ask it where you keep your iron and it will tell you or go get it. The memory persists across play sessions, so your base layout is remembered even after relogging.

Waypoint management is also available — you can ask your companion to mark a chest location, look up what it knows about your storage, or clear outdated entries.

---

💰 Budget Controls

New /player2npc budget commands let you control how much your companions are allowed to spend on AI within a rolling time window. Set a limit, and the mod enforces it automatically — no surprises on your Joules balance after a long session.

/player2npc budget status — shows your current limits, how many AI calls have been made this window, and your cached Joules balance.

/player2npc budget reset — clears the current call window immediately so AI requests are allowed again. Useful if you hit a limit and want to keep going.

/player2npc budget window <minutes> — sets how long the rolling window lasts before call counts reset. For example, window 60 gives you a fresh allowance every hour.

/player2npc budget soft <calls> — sets a soft call limit. When hit, the mod may route to a lighter AI model instead of stopping entirely. Set to 0 to disable.

/player2npc budget hard <calls> — sets a hard call limit. When hit, AI calls stop completely until the window resets. Set to 0 to disable.

/player2npc budget joules_soft <amount> — same idea as the soft call limit, but based on your Joules balance rather than call count. Triggers when your balance drops to or below this amount.

/player2npc budget joules_hard <amount> — hard stop based on Joules balance. AI calls are blocked until your balance recovers above this threshold.

/player2npc budget joules_refresh <seconds> — controls how often your Joules balance is re-checked from the Player2 platform. Defaults are conservative to avoid unnecessary API calls.

On servers where the owner pays for AI costs rather than individual players, operators can manage limits from the engine side with /playerengine player2 budget.

---

🤖 AI & Power User Updates

Smarter Tool Retrieval

The AI now uses on-device semantic search to pick the right tool for a given request, rather than loading the full command list into every prompt. This keeps things efficient as the available command set grows, and generally makes tool selection more accurate.

Learning From Mistakes

When the AI retrieves the wrong tasks, it automatically retries with a rephrased query and records the failure as a learning signal. Over time, the system builds up accurate aliases for your usage patterns so the same retrieval miss doesn't repeat. Currently off by default. Turn on in /playerengine/server_player2.json (Deepsearch and related settings)

Model Tier Routing

AI calls are now automatically routed to the appropriate model tier based on the task. Heavy planning calls can use a stronger model; lightweight follow-ups use a cheaper one. This happens automatically based on your profile and patron tier — nothing to configure.

Mod Intelligence

The companion now scans the mods installed on your server and builds a background knowledge base of their items, blocks, and entities. This runs quietly after world load and improves over time. Companions with this knowledge can reason about modded content more accurately — recognizing what an unfamiliar modded item does rather than treating it as completely unknown.

You can check the status or trigger a rebuild with /playerengine capability status. See the enrichment guide for setup and tuning details.

Budget Guards

The budget enforcement runs at the engine level on every AI call — in singleplayer and on servers alike. Patrons with a named profile can designate it as a fallback that the engine automatically switches to when a soft limit is hit, so usage can step down to a cheaper model before stopping entirely. On dedicated servers where each player pays for their own usage, limits track per-player rather than globally.

---

🐛 Bug Fixes

Crash on World Load With Certain Mods

An internal hook was triggering crashes when loading into worlds that contained mods with complex item behavior — notably SilentGear and Construction Sticks in modpacks like All the Mods 10. The game would crash before gameplay even started. Fixed.

Multiple Companions Talking Over Each Other

When more than one companion was active, they would respond simultaneously — a companion would start speaking before the previous one had finished. Companions now wait for the current speaker's audio to complete before queuing their own response. Fixed.

Companion Wandering Instead of Moving to a Target

An overflow bug in the companion's movement system caused it to treat any block approach as already complete, so it would just wander instead of walking up to what it needed to reach. This affected any task involving moving to a specific block — chests, crafting tables, resource nodes. Fixed.

---

⚠️ Known Issues

gamer command (autonomous "beat the game" run) is temporarily disabled. It will return in a future update once outstanding issues with its execution are resolved.

Voice transcription (STT) may produce empty results for some users. If your companion isn't responding to voice input but responds fine to typed chat, check latest.log for [STT] entries — this will help us narrow down the cause.

Multiplayer update!

• Dedicated servers support! (See note at bottom)
• Local hosted servers
• LAN play (untested probably works I don't have a way to test)

Server configs/commands!

• Control data storage
• Control AI spawn limits
• Users can control whose bots can respond to your prompts
• Users can control who can talk to your bots
• Everyone gets their own bots, no need to share! (admins set your spawn/storage limits)
• Want to pay for your friends/family bot usage on a small server? Set payer mode to owner! (Not recommended for large/public groups)

New Features!

• AI can place and read signs. (Weaker AI models may struggle. Hanging signs currently unsupported. Trying to place hanging shouldn't crash but don't recommend trying.)
• AI can now get in and out of boats (Most reliable method is to get it to follow you first)
• Call by name: Default/recommended setting especially multiplayer. This setting will make bots ignore messages not addressed to them. Simply use the AI's full name anywhere in your message. You can even address multiple AI in one message. Name should be their full chat name or as it's written over their head. Names are case INSENSITIVE, eLLie works the same as Ellie. Will prefer your own bot in the case of duplicates. You can also address another player's copy directly by writing 's in your message. Ex: "Hey Ellie, what do you think of Zardov's Ellie?" This would result in both bots recieving the message.

*Note: Server admins should use /playerengine player2 dedicated true when not hosting on a local machine. Since most remote hosting services do not support running the player2 app on their servers and cannot auth the web api for you. This switches the server to hand off api calls to clients. And is honestly the recommended setup in most multiplayer cases since it also improves api request que times for multiplayer. So you may even want to use this setting on a local hosted server. WILL NOT HELP SINGLEPLAYER. You won't get any benefits to que time if it's just one player.

Multiplayer update!

• Dedicated servers support! (See note at bottom)
• Local hosted servers
• LAN play (untested probably works I don't have a way to test)

Server configs/commands!

• Control data storage
• Control AI spawn limits
• Users can control whose bots can respond to your prompts
• Users can control who can talk to your bots
• Everyone gets their own bots, no need to share! (admins set your spawn/storage limits)
• Want to pay for your friends/family bot usage on a small server? Set payer mode to owner! (Not recommended for large/public groups)

New Features!

• AI can place and read signs. (Weaker AI models may struggle. Hanging signs currently unsupported. Trying to place hanging shouldn't crash but don't recommend trying.)
• AI can now get in and out of boats (Most reliable method is to get it to follow you first)
• Call by name: Default/recommended setting especially multiplayer. This setting will make bots ignore messages not addressed to them. Simply use the AI's full name anywhere in your message. You can even address multiple AI in one message. Name should be their full chat name or as it's written over their head. Names are case INSENSITIVE, eLLie works the same as Ellie. Will prefer your own bot in the case of duplicates. You can also address another player's copy directly by writing 's in your message. Ex: "Hey Ellie, what do you think of Zardov's Ellie?" This would result in both bots recieving the message.

*Note: Server admins should use /playerengine player2 dedicated true when not hosting on a local machine. Since most remote hosting services do not support running the player2 app on their servers and cannot auth the web api for you. This switches the server to hand off api calls to clients. And is honestly the recommended setup in most multiplayer cases since it also improves api request que times for multiplayer. So you may even want to use this setting on a local hosted server. WILL NOT HELP SINGLEPLAYER. You won't get any benefits to que time if it's just one player.

Multiplayer update!

• Dedicated servers support! (See note at bottom)
• Local hosted servers
• LAN play (untested probably works I don't have a way to test)

Server configs/commands!

• Control data storage
• Control AI spawn limits
• Users can control whose bots can respond to your prompts
• Users can control who can talk to your bots
• Everyone gets their own bots, no need to share! (admins set your spawn/storage limits)
• Want to pay for your friends/family bot usage on a small server? Set payer mode to owner! (Not recommended for large/public groups)

New Features!

• AI can place and read signs. (Weaker AI models may struggle. Hanging signs currently unsupported. Trying to place hanging shouldn't crash but don't recommend trying.)
• AI can now get in and out of boats (Most reliable method is to get it to follow you first)
• Call by name: Default/recommended setting especially multiplayer. This setting will make bots ignore messages not addressed to them. Simply use the AI's full name anywhere in your message. You can even address multiple AI in one message. Name should be their full chat name or as it's written over their head. Names are case INSENSITIVE, eLLie works the same as Ellie. Will prefer your own bot in the case of duplicates. You can also address another player's copy directly by writing 's in your message. Ex: "Hey Ellie, what do you think of Zardov's Ellie?" This would result in both bots recieving the message.

*Note: Server admins should use /playerengine player2 dedicated true when not hosting on a local machine. Since most remote hosting services do not support running the player2 app on their servers and cannot auth the web api for you. This switches the server to hand off api calls to clients. And is honestly the recommended setup in most multiplayer cases since it also improves api request que times for multiplayer. So you may even want to use this setting on a local hosted server. WILL NOT HELP SINGLEPLAYER. You won't get any benefits to que time if it's just one player.

Multiplayer update!

• Dedicated servers support! (See note at bottom)
• Local hosted servers
• LAN play (untested probably works I don't have a way to test)

Server configs/commands!

• Control data storage
• Control AI spawn limits
• Users can control whose bots can respond to your prompts
• Users can control who can talk to your bots
• Everyone gets their own bots, no need to share! (admins set your spawn/storage limits)
• Want to pay for your friends/family bot usage on a small server? Set payer mode to owner! (Not recommended for large/public groups)

New Features!

• AI can place and read signs. (Weaker AI models may struggle. Hanging signs currently unsupported. Trying to place hanging shouldn't crash but don't recommend trying.)
• AI can now get in and out of boats (Most reliable method is to get it to follow you first)
• Call by name: Default/recommended setting especially multiplayer. This setting will make bots ignore messages not addressed to them. Simply use the AI's full name anywhere in your message. You can even address multiple AI in one message. Name should be their full chat name or as it's written over their head. Names are case INSENSITIVE, eLLie works the same as Ellie. Will prefer your own bot in the case of duplicates. You can also address another player's copy directly by writing 's in your message. Ex: "Hey Ellie, what do you think of Zardov's Ellie?" This would result in both bots recieving the message.

*Note: Server admins should use /playerengine player2 dedicated true when not hosting on a local machine. Since most remote hosting services do not support running the player2 app on their servers and cannot auth the web api for you. This switches the server to hand off api calls to clients. And is honestly the recommended setup in most multiplayer cases since it also improves api request que times for multiplayer. So you may even want to use this setting on a local hosted server. WILL NOT HELP SINGLEPLAYER. You won't get any benefits to que time if it's just one player.