Pi Coding Agent - Revision history

Davrot@uni-bremen.de: /* Installing Ollama: */

2026-09-15T14:48:46Z

Installing Ollama:

← Older revision		Revision as of 14:48, 15 September 2026
Line 235:		Line 235:

	still as root<syntaxhighlight lang="bash">		still as root<syntaxhighlight lang="bash">
			systemctl daemon-reload
	systemctl enable ollama		systemctl enable ollama
	systemctl start ollama		systemctl start ollama

Davrot@uni-bremen.de: /* 2019-era AMD EPYC 7282 with 256GB (UD-IQ4_XS) */

2026-08-30T02:29:46Z

2019-era AMD EPYC 7282 with 256GB (UD-IQ4_XS)

@@ Line 317: / Line 317: @@
 "predicted_per_second":3.56101122618500<syntaxhighlight lang="yaml">
 services:
    llama-cpp:

Davrot@uni-bremen.de: /* Qwen 3.8 Flash Next on CPU */

2026-08-30T02:29:25Z

Qwen 3.8 Flash Next on CPU

@@ Line 309: / Line 309: @@
      "max_tokens": 64
    }'
 </syntaxhighlight>

Davrot@uni-bremen.de: /* Installing Ollama: */

2026-08-30T01:20:25Z

Installing Ollama:

@@ Line 241: / Line 241: @@
 </syntaxhighlight>as normal user (or root):<syntaxhighlight lang="bash">ollama pull smtek/Qwen3.8-27B:IQ2_M-256k</syntaxhighlight>Now it should run. For testing:<syntaxhighlight lang="bash">
 ollama run smtek/Qwen3.8-27B:IQ2_M-256k
 </syntaxhighlight>

Davrot@uni-bremen.de: /* Running Qwen 3.8 locally */

2026-08-21T16:37:40Z

Running Qwen 3.8 locally

← Older revision		Revision as of 16:37, 21 August 2026
Line 164:		Line 164:
	},		},
	"contextWindow": 196608,		"contextWindow": 196608,
	"maxTokens": ~~196608~~		"maxTokens": 16384
	}		}
	]		]
Line 204:		Line 204:
	"httpIdleTimeoutMs": 1800000		"httpIdleTimeoutMs": 1800000
	}		}
			</syntaxhighlight>[hermes-memory-config.json]<syntaxhighlight lang="json">
			{
			"llmModelOverride": "smtek/Qwen3.8-27B:IQ2_M-256k",
			"llmThinkingOverride": "off",
			"consolidationTimeoutMs": 600000,
			"temperature": 0.1
			}

	</syntaxhighlight>		</syntaxhighlight>

Davrot@uni-bremen.de: /* Running Qwen 3.8 locally */

2026-08-19T15:34:10Z

Running Qwen 3.8 locally

← Older revision		Revision as of 15:34, 19 August 2026
Line 204:		Line 204:
	"httpIdleTimeoutMs": 1800000		"httpIdleTimeoutMs": 1800000
	}		}
	</syntaxhighlight>~~Installing Ollama:~~		</syntaxhighlight>

			=== Installing Ollama: ===
	as root:<syntaxhighlight lang="bash">		as root:<syntaxhighlight lang="bash">
	curl -fsSL https://ollama.com/install.sh \| sh		curl -fsSL https://ollama.com/install.sh \| sh
Line 230:		Line 231:
	systemctl status ollama		systemctl status ollama

	</syntaxhighlight>as normal user (or root):<syntaxhighlight lang="bash">ollama pull smtek/Qwen3.8-27B:IQ2_M-256k</syntaxhighlight>Now it should run:<syntaxhighlight lang="bash">		</syntaxhighlight>as normal user (or root):<syntaxhighlight lang="bash">ollama pull smtek/Qwen3.8-27B:IQ2_M-256k</syntaxhighlight>Now it should run. For testing:<syntaxhighlight lang="bash">
	ollama run smtek/Qwen3.8-27B:IQ2_M-256k		ollama run smtek/Qwen3.8-27B:IQ2_M-256k
	</syntaxhighlight>		</syntaxhighlight>

Davrot@uni-bremen.de: /* Running Qwen 3.8 locally */

2026-08-19T15:33:27Z

Running Qwen 3.8 locally

← Older revision		Revision as of 15:33, 19 August 2026
Line 146:		Line 146:


	[models.json]<syntaxhighlight lang="json">		[models.json] obviously you want to change your '''baseUrl!'''<syntaxhighlight lang="json">
	{		{
	"providers": {		"providers": {

Davrot@uni-bremen.de: /* Running Qwen 3.8 locally */

2026-08-19T15:32:54Z

Running Qwen 3.8 locally

← Older revision		Revision as of 15:32, 19 August 2026
Line 142:		Line 142:
	ollama: smtek/Qwen3.8-27B:IQ2_M-256k on 24GB RTX3090		ollama: smtek/Qwen3.8-27B:IQ2_M-256k on 24GB RTX3090

	~~<nowiki>~~https://ollama.com/smtek/Qwen3.8-27B~~</nowiki>~~		Quants from Q2 up to Q5 from Unsloth K_M and UD_K_XL and builds for 12, 16 and 24 GB: https://ollama.com/smtek/Qwen3.8-27B


	~~Quants from Q2 up to Q5 from Unsloth K_M and UD_K_XL and builds for 12, 16 and 24 GB~~

	[models.json]<syntaxhighlight lang="json">		[models.json]<syntaxhighlight lang="json">

Davrot@uni-bremen.de: /* Pi Coding Agent Error : 500 unexpected message role developer */

2026-08-19T15:32:21Z

Pi Coding Agent Error : 500 unexpected message role developer

← Older revision		Revision as of 15:32, 19 August 2026
Line 138:		Line 138:
	}		}
	}</syntaxhighlight>		}</syntaxhighlight>

			== Running Qwen 3.8 locally ==
			ollama: smtek/Qwen3.8-27B:IQ2_M-256k on 24GB RTX3090

			<nowiki>https://ollama.com/smtek/Qwen3.8-27B</nowiki>

			Quants from Q2 up to Q5 from Unsloth K_M and UD_K_XL and builds for 12, 16 and 24 GB

			[models.json]<syntaxhighlight lang="json">
			{
			"providers": {
			"ollama": {
			"baseUrl": "http://gate2.neuro.uni-bremen.de:11434/v1",
			"api": "openai-completions",
			"apiKey": "ollama",
			"compat": {
			"supportsDeveloperRole": false
			},
			"models": [
			{
			"id": "smtek/Qwen3.8-27B:IQ2_M-256k",
			"reasoning": true,
			"compat": {
			"thinkingFormat": "qwen"
			},
			"contextWindow": 196608,
			"maxTokens": 196608
			}
			]
			}
			}
			}
			</syntaxhighlight>[settings.json]<syntaxhighlight lang="json">
			{
			"packages": [
			"npm:pi-subagents",
			"npm:pi-mcp-adapter",
			"npm:@narumitw/pi-lsp",
			"npm:pi-agent-browser-native",
			"npm:pi-web-access",
			{
			"source": "npm:pi-agent-extensions",
			"extensions": [
			"!/whimsical/"
			]
			},
			"npm:pi-codex-goal",
			"npm:context-mode",
			"npm:pi-hermes-memory"
			],
			"lastChangelogVersion": "0.84.2",
			"theme": "dark",
			"defaultProvider": "ollama",
			"defaultModel": "smtek/Qwen3.8-27B:IQ2_M-256k",
			"retry": {
			"enabled": true,
			"maxRetries": 5,
			"baseDelayMs": 5000,
			"provider": {
			"timeoutMs": 3600000,
			"maxRetries": 0,
			"maxRetryDelayMs": 60000
			}
			},
			"httpIdleTimeoutMs": 1800000
			}
			</syntaxhighlight>Installing Ollama:

			as root:<syntaxhighlight lang="bash">
			curl -fsSL https://ollama.com/install.sh \| sh
			mkdir /etc/systemd/system/ollama.service.d
			</syntaxhighlight>Optional:

			Setting the model directory (i.e. on a SSD) and making OLLAMA via port 11434 available for other computers

			[/etc/systemd/system/ollama.service.d/override.conf]<syntaxhighlight lang="bash">
			[Service]
			Environment="OLLAMA_MODELS=/ollama"
			Environment="OLLAMA_HOST=0.0.0.0:11434"

			</syntaxhighlight>still as root:<syntaxhighlight lang="bash">
			mkdir -p /ollama
			chown -R ollama:ollama /ollama
			chmod -R u+rwX /ollama

			</syntaxhighlight>END OPTIONAL

			still as root<syntaxhighlight lang="bash">
			systemctl enable ollama
			systemctl start ollama
			systemctl status ollama

			</syntaxhighlight>as normal user (or root):<syntaxhighlight lang="bash">ollama pull smtek/Qwen3.8-27B:IQ2_M-256k</syntaxhighlight>Now it should run:<syntaxhighlight lang="bash">
			ollama run smtek/Qwen3.8-27B:IQ2_M-256k
			</syntaxhighlight>

Davrot@uni-bremen.de: /* Multi-Agent Workflows with pi-subagents */

2026-08-14T18:33:13Z

Multi-Agent Workflows with pi-subagents

← Older revision		Revision as of 18:33, 14 August 2026
Line 99:		Line 99:
	* '''Include <code>AGENTS.md</code> in your opening prompt:''' Start new sessions with a standard directive like <code>"Read AGENTS.md first to load project context."</code> This instantly grounds Pi in your team's workflow without manual setup each time.		* '''Include <code>AGENTS.md</code> in your opening prompt:''' Start new sessions with a standard directive like <code>"Read AGENTS.md first to load project context."</code> This instantly grounds Pi in your team's workflow without manual setup each time.
	* '''Persist project knowledge:''' Instruct Pi to update <code>AGENTS.md</code> whenever core architecture, technology choices, or coding conventions evolve. Treating <code>AGENTS.md</code> as living documentation ensures future agent sessions stay aligned without repeated prompt engineering.		* '''Persist project knowledge:''' Instruct Pi to update <code>AGENTS.md</code> whenever core architecture, technology choices, or coding conventions evolve. Treating <code>AGENTS.md</code> as living documentation ensures future agent sessions stay aligned without repeated prompt engineering.

			== Pi Coding Agent Error : 500 unexpected message role developer ==
			If you see this<syntaxhighlight lang="json">
			Error: 500: {"message":"unexpected message role \"developer\"","type":"api_error","param":null,"code":null}
			</syntaxhighlight>you need to add "supportsDeveloperRole": false to /root/.pi/agent/models.json : <syntaxhighlight lang="json">{
			"providers": {
			"ollama": {
			"baseUrl": "http://localhost:11434/v1",
			"api": "openai-completions",
			"apiKey": "ollama",
			"compat": {
			"supportsDeveloperRole": false
			},
			"models": [
			{
			"id": "qwen3-coder-next:latest",
			"reasoning": false,
			"compat": {
			"thinkingFormat": "qwen"
			},
			"contextWindow": 262144,
			"maxTokens": 262144
			},
			{
			"id": "qwen3.8:latest",
			"reasoning": true,
			"compat": {
			"thinkingFormat": "qwen"
			},
			"contextWindow": 262144,
			"maxTokens": 262144,
			"options": {
			"reasoning_effort": "xhigh"
			}
			}
			]
			}
			}
			}</syntaxhighlight>