feat: context window from model endpoints + percentage-based auto-compact

ModelInfo (kernel contract):
- Add contextWindow?: number field

OpenAI-stream listModels:
- Parse contextWindow from common field names (context_length,
  context_window, max_context_length, max_tokens)

Transport-contract:
- ModelsResponse: add optional modelInfo map (model name → { contextWindow? })
- Add ModelMetadata type
- Rename CompactThresholdResponse → CompactPercentResponse
- Rename SetCompactThresholdRequest → SetCompactPercentRequest

Credential store:
- Add getModelInfo(modelName) method — resolves full ModelInfo
  (including contextWindow) for a <credential>/<model> string

Transport-http:
- GET /models now includes modelInfo with contextWindow per model
- Rename compact-threshold endpoints → compact-percent

Session-orchestrator:
- Auto-compact now uses contextSize (not overcounted usage.inputTokens)
  compared against contextWindow * (percent / 100)
- Default percent: 85 (was flat 350000)
- resolveModelInfo dep added to look up contextWindow
- Passes modelName from the settled turn to the compaction service

Conversation store:
- Rename getCompactThreshold/setCompactThreshold → getCompactPercent/setCompactPercent
- compactThresholdKey → compact-percent key

17 files changed, 184 insertions, 87 deletions

diff --git a/backend-to-fe-handoff.md b/backend-to-fe-handoff.md
index 1e9ffb2..dcaaa51 100644
--- a/backend-to-fe-handoff.md
+++ b/backend-to-fe-handoff.md
@@ -45,7 +45,7 @@ after `runTurn` returns — backward compatible.
 |---|---|---|
 | `POST` | `/chat` | Stream a turn (NDJSON response, `X-Conversation-Id` header) |
 | `POST` | `/chat/warm` | Cache-warm probe |
-| `GET` | `/models` | List available models |
+| `GET` | `/models` | Model catalog (now includes `modelInfo` with `contextWindow` per model) |
 | `GET` | `/conversations` | List conversations (`?q=` prefix filter, `?status=active,idle` status filter) |
 | `GET` | `/conversations/:id` | Conversation history (`?sinceSeq=`, `?beforeSeq=`, `?limit=` windowing) |
 | `GET` | `/conversations/:id/metrics` | Per-turn metrics (tokens, timing) |
@@ -55,8 +55,8 @@ after `runTurn` returns — backward compatible.
 | `GET` | `/conversations/:id/reasoning-effort` | Per-conversation reasoning effort |
 | `PUT` | `/conversations/:id/reasoning-effort` | Set reasoning effort |
 | `GET` | `/conversations/:id/lsp` | LSP server status |
-| `GET` | `/conversations/:id/compact-threshold` | Auto-compact threshold (0=manual, null=default 350k) |
-| `PUT` | `/conversations/:id/compact-threshold` | Set auto-compact threshold |
+| `GET` | `/conversations/:id/compact-percent` | Auto-compact percent (0=manual, null=default 85%) |
+| `PUT` | `/conversations/:id/compact-percent` | Set auto-compact percent |
 | `GET` | `/conversations/:id/title` | Read conversation title |
 | `PUT` | `/conversations/:id/title` | Set conversation title |
 | `POST` | `/conversations/:id/close` | Close tab (abort turn + mark `closed`) |
@@ -119,13 +119,13 @@ interface CompactResponse {
   messagesKept: number;
 }
 
-interface CompactThresholdResponse {
+interface CompactPercentResponse {
   conversationId: string;
-  threshold: number;  // 0 = manual; null = default 350000
+  percent: number;  // 0 = manual; null = default 85
 }
 
-interface SetCompactThresholdRequest {
-  threshold: number;
+interface SetCompactPercentRequest {
+  percent: number;
 }
 ```
 
diff --git a/frontend-compaction-handoff.md b/frontend-compaction-handoff.md
index 6979c29..572b25f 100644
--- a/frontend-compaction-handoff.md
+++ b/frontend-compaction-handoff.md
@@ -11,7 +11,7 @@ Creates a linked chain of archives you can walk backward.
 Two modes:
 - **Manual**: `POST /conversations/:id/compact` — triggers immediately.
 - **Automatic**: after each turn settles, the backend checks if the last turn's
-  input tokens exceeded the per-conversation `compactThreshold` (default 350000).
+  input tokens exceeded the per-conversation `compactThreshold` (default 85).
   If so, compaction runs automatically (fire-and-forget, non-blocking).
 
 ## How compaction works — non-destructive, chained
@@ -84,13 +84,13 @@ export interface CompactResponse {
   readonly messagesKept: number;
 }
 
-export interface CompactThresholdResponse {
+export interface CompactPercentResponse {
   readonly conversationId: string;
-  readonly threshold: number;  // 0 = manual only; null = default 350000
+  readonly percent: number;  // 0 = manual only; null = default 85
 }
 
-export interface SetCompactThresholdRequest {
-  readonly threshold: number;
+export interface SetCompactPercentRequest {
+  readonly percent: number;
 }
 ```
 
@@ -107,24 +107,24 @@ Triggers compaction on demand. Optional JSON body:
 The conversation ID in the response is the same as the request — the ID doesn't
 change. The FE should reload the conversation history.
 
-409: `{ error: string }` — conversation is generating, too short, threshold not exceeded, etc.
+409: `{ error: string }` — conversation is generating, too short, percent not exceeded, etc.
 503: compaction service not available.
 
-## `GET /conversations/:id/compact-threshold` — read threshold
+## `GET /conversations/:id/compact-percent` — read percent
 
-200: `CompactThresholdResponse { conversationId, threshold }`
-- `threshold: 0` — auto-compact explicitly disabled (manual only).
-- `threshold: null` (not stored) — **default: 350000** (350k tokens). The FE
-  should display 350000 as the default value in the settings UI.
+200: `CompactPercentResponse { conversationId, percent }`
+- `percent: 0` — auto-compact explicitly disabled (manual only).
+- `percent: null` (not stored) — **default: 85** (85% tokens). The FE
+  should display 85 as the default value in the settings UI.
 - Any positive number — auto-compact triggers when the last turn's input tokens
   exceed this value.
 
-## `PUT /conversations/:id/compact-threshold` — set threshold
+## `PUT /conversations/:id/compact-percent` — set percent
 
-Body: `SetCompactThresholdRequest { threshold: number }`
+Body: `SetCompactPercentRequest { percent: number }`
 - `0` explicitly disables auto-compact.
-- Any positive number sets the trigger threshold.
-- To "reset to default", set it to 350000.
+- Any positive number sets the trigger percent.
+- To "reset to default", set it to 85.
 
 ## `conversation.compacted` WS message
 
@@ -141,9 +141,9 @@ No tab switching needed — the ID is the same.
    Show a loading indicator while waiting. On success, reload the conversation
    history (same ID — just re-fetch).
 
-2. **Settings UI** for compact threshold: `PUT /conversations/:id/compact-threshold`
-   with `{ threshold: number }`. A number input (0 = manual only, default 350000).
-   Read the current value via `GET /conversations/:id/compact-threshold`.
+2. **Settings UI** for compact percent: `PUT /conversations/:id/compact-percent`
+   with `{ percent: number }`. A number input (0 = manual only, default 85).
+   Read the current value via `GET /conversations/:id/compact-percent`.
 
 3. **Handle `conversation.compacted` WS messages**: reload the conversation
    history via `GET /conversations/:id` (same ID, no tab switch).
diff --git a/packages/conversation-store/src/keys.ts b/packages/conversation-store/src/keys.ts
index a609f82..18c683f 100644
--- a/packages/conversation-store/src/keys.ts
+++ b/packages/conversation-store/src/keys.ts
@@ -55,7 +55,7 @@ export function reasoningEffortKey(conversationId: string): string {
 }
 
 export function compactThresholdKey(conversationId: string): string {
-	return `conv:${conversationId}:compact-threshold`;
+	return `conv:${conversationId}:compact-percent`;
 }
 
 export function metaKey(conversationId: string): string {
diff --git a/packages/conversation-store/src/store.ts b/packages/conversation-store/src/store.ts
index 700be1e..8c287ae 100644
--- a/packages/conversation-store/src/store.ts
+++ b/packages/conversation-store/src/store.ts
@@ -105,10 +105,10 @@ export interface ConversationStore {
 	 * non-destructively before replacing it with a summary.
 	 */
 	readonly forkHistory: (sourceId: string, targetId: string) => Promise<void>;
-	/** Get the compact threshold (token count, 0 = manual only), or null if unset. */
-	readonly getCompactThreshold: (conversationId: string) => Promise<number | null>;
-	/** Set the compact threshold (token count, 0 = manual only). */
-	readonly setCompactThreshold: (conversationId: string, threshold: number) => Promise<void>;
+	/** Get the compact percent (0-100, 0 = manual only), or null if unset. */
+	readonly getCompactPercent: (conversationId: string) => Promise<number | null>;
+	/** Set the compact percent (0-100, 0 = manual only). */
+	readonly setCompactPercent: (conversationId: string, percent: number) => Promise<void>;
 	/**
 	 * Set the `compactedFrom` field on a conversation's metadata, pointing to
 	 * the archive conversation that holds the pre-compaction history.
@@ -620,17 +620,17 @@ export function createConversationStore(
 			if (effort !== null) await storage.set(reasoningEffortKey(targetId), effort);
 		},
 
-		async getCompactThreshold(conversationId) {
+		async getCompactPercent(conversationId) {
 			const raw = await storage.get(compactThresholdKey(conversationId));
 			if (raw === null) return null;
 			const n = Number.parseInt(raw, 10);
 			return Number.isNaN(n) ? null : n;
 		},
 
-		async setCompactThreshold(conversationId, threshold) {
-			await storage.set(compactThresholdKey(conversationId), String(threshold));
+		async setCompactPercent(conversationId, percent) {
+			await storage.set(compactThresholdKey(conversationId), String(percent));
 			if (logger !== undefined) {
-				logger.debug("compact-threshold set", { conversationId, threshold });
+				logger.debug("compact-percent set", { conversationId, percent });
 			}
 		},
 
diff --git a/packages/credential-store/src/registry.ts b/packages/credential-store/src/registry.ts
index c1ba771..89d8084 100644
--- a/packages/credential-store/src/registry.ts
+++ b/packages/credential-store/src/registry.ts
@@ -1,4 +1,4 @@
-import type { ProviderContract } from "@dispatch/kernel";
+import type { ModelInfo, ProviderContract } from "@dispatch/kernel";
 
 export interface Credential {
 	readonly name: string;
@@ -24,6 +24,13 @@ export interface CredentialStore {
 	 * missing or has no listModels.
 	 */
 	listCatalog(): Promise<readonly string[]>;
+
+	/**
+	 * Returns the full `ModelInfo` for a `<credentialName>/<model>` string, or
+	 * undefined if unknown. Caches the result of `listModels` per credential.
+	 * Used to look up `contextWindow` for auto-compaction.
+	 */
+	getModelInfo(modelName: string): Promise<ModelInfo | undefined>;
 }
 
 export interface CredentialStoreDeps {
@@ -76,5 +83,22 @@ export function createCredentialStore(deps: CredentialStoreDeps): CredentialStor
 
 			return results;
 		},
+
+		async getModelInfo(modelName: string): Promise<ModelInfo | undefined> {
+			const slashIndex = modelName.indexOf("/");
+			if (slashIndex === -1) return undefined;
+			const credentialName = modelName.slice(0, slashIndex);
+			const modelId = modelName.slice(slashIndex + 1);
+			if (!modelId) return undefined;
+
+			const providerId = credentialMap.get(credentialName);
+			if (providerId === undefined) return undefined;
+
+			const provider = deps.getProvider(providerId);
+			if (provider?.listModels === undefined) return undefined;
+
+			const models = await provider.listModels();
+			return models.find((m) => m.id === modelId);
+		},
 	};
 }
diff --git a/packages/kernel/src/contracts/provider.ts b/packages/kernel/src/contracts/provider.ts
index 7f920c5..52d853b 100644
--- a/packages/kernel/src/contracts/provider.ts
+++ b/packages/kernel/src/contracts/provider.ts
@@ -112,6 +112,8 @@ export interface ProviderStreamOptions {
 export interface ModelInfo {
 	readonly id: string;
 	readonly displayName?: string;
+	/** The model's max context window in tokens (e.g. 200000). Optional — providers that don't report it leave it undefined. */
+	readonly contextWindow?: number;
 }
 
 /**
diff --git a/packages/openai-stream/src/listModels.ts b/packages/openai-stream/src/listModels.ts
index 3f783f0..8a76c67 100644
--- a/packages/openai-stream/src/listModels.ts
+++ b/packages/openai-stream/src/listModels.ts
@@ -13,6 +13,10 @@ import type { FetchLike } from "@dispatch/trace-replay";
 
 interface OpenAIModelEntry {
 	readonly id: string;
+	readonly context_length?: number;
+	readonly context_window?: number;
+	readonly max_context_length?: number;
+	readonly max_tokens?: number;
 }
 
 interface OpenAIModelListResponse {
@@ -21,10 +25,18 @@ interface OpenAIModelListResponse {
 
 /**
  * Pure mapping: raw OpenAI-compatible model list → ModelInfo[].
+ * Extracts `contextWindow` from common field names (providers vary).
  * Extracted for direct unit testing with no I/O.
  */
 export function parseModelList(data: readonly OpenAIModelEntry[]): readonly ModelInfo[] {
-	return data.map((entry) => ({ id: entry.id }));
+	return data.map((entry) => {
+		const contextWindow =
+			entry.context_length ?? entry.context_window ?? entry.max_context_length ?? entry.max_tokens;
+		return {
+			id: entry.id,
+			...(contextWindow !== undefined ? { contextWindow } : {}),
+		};
+	});
 }
 
 export interface ListModelsConfig {
diff --git a/packages/session-orchestrator/src/extension.ts b/packages/session-orchestrator/src/extension.ts
index ae90902..fca8ddb 100644
--- a/packages/session-orchestrator/src/extension.ts
+++ b/packages/session-orchestrator/src/extension.ts
@@ -53,6 +53,10 @@ export function activate(host: HostAPI): void {
 			const provider = host.getProviders().get(r.providerId);
 			return provider ? { provider, model: r.model } : undefined;
 		},
+		resolveModelInfo: async (modelName: string) => {
+			const store = host.getService(credentialStoreHandle);
+			return store.getModelInfo(modelName);
+		},
 		applyToolsFilter: (assembly) => host.applyFilters(toolsFilter, assembly),
 		runTurn,
 		logger: host.logger,
@@ -116,6 +120,10 @@ export function activate(host: HostAPI): void {
 				const provider = host.getProviders().get(r.providerId);
 				return provider ? { provider, model: r.model } : undefined;
 			},
+			resolveModelInfo: async (modelName: string) => {
+				const store = host.getService(credentialStoreHandle);
+				return store.getModelInfo(modelName);
+			},
 			applyToolsFilter: (assembly) => host.applyFilters(toolsFilter, assembly),
 			runTurn,
 			logger: host.logger,
diff --git a/packages/session-orchestrator/src/orchestrator.test.ts b/packages/session-orchestrator/src/orchestrator.test.ts
index 012ac5c..ce134f4 100644
--- a/packages/session-orchestrator/src/orchestrator.test.ts
+++ b/packages/session-orchestrator/src/orchestrator.test.ts
@@ -91,10 +91,10 @@ function createInMemoryStore(): ConversationStore & {
 		},
 		async setConversationStatus() {},
 		async replaceHistory() {},
-		async getCompactThreshold() {
+		async getCompactPercent() {
 			return null;
 		},
-		async setCompactThreshold() {},
+		async setCompactPercent() {},
 		async forkHistory() {},
 		async setCompactedFrom() {},
 	};
@@ -548,10 +548,10 @@ describe("turn-sealed event", () => {
 			},
 			async setConversationStatus() {},
 			async replaceHistory() {},
-			async getCompactThreshold() {
+			async getCompactPercent() {
 				return null;
 			},
-			async setCompactThreshold() {},
+			async setCompactPercent() {},
 			async forkHistory() {},
 			async setCompactedFrom() {},
 		};
@@ -621,10 +621,10 @@ describe("turn-sealed event", () => {
 			},
 			async setConversationStatus() {},
 			async replaceHistory() {},
-			async getCompactThreshold() {
+			async getCompactPercent() {
 				return null;
 			},
-			async setCompactThreshold() {},
+			async setCompactPercent() {},
 			async forkHistory() {},
 			async setCompactedFrom() {},
 		};
@@ -983,10 +983,10 @@ describe("turn metrics persistence", () => {
 			},
 			async setConversationStatus() {},
 			async replaceHistory() {},
-			async getCompactThreshold() {
+			async getCompactPercent() {
 				return null;
 			},
-			async setCompactThreshold() {},
+			async setCompactPercent() {},
 			async forkHistory() {},
 			async setCompactedFrom() {},
 		};
diff --git a/packages/session-orchestrator/src/orchestrator.ts b/packages/session-orchestrator/src/orchestrator.ts
index 22a5c11..adf5680 100644
--- a/packages/session-orchestrator/src/orchestrator.ts
+++ b/packages/session-orchestrator/src/orchestrator.ts
@@ -6,6 +6,7 @@ import type {
 	ConversationStatus,
 	EventHookDescriptor,
 	Logger,
+	ModelInfo,
 	ProviderContract,
 	ProviderEvent,
 	ProviderStreamOptions,
@@ -248,6 +249,12 @@ export interface SessionOrchestratorDeps {
 	readonly resolveModel?: (
 		modelName: string,
 	) => { provider: ProviderContract; model: string } | undefined;
+	/**
+	 * Resolve full `ModelInfo` (including `contextWindow`) for a model name.
+	 * Used by the compaction service to calculate the auto-compact threshold
+	 * as a percentage of the context window.
+	 */
+	readonly resolveModelInfo?: (modelName: string) => Promise<ModelInfo | undefined>;
 	readonly runTurn: (input: RunTurnInput) => Promise<RunTurnResult>;
 	/**
 	 * Lazily resolves the message-queue service (the steering queue), or
@@ -519,7 +526,12 @@ export function createSessionOrchestrator(
 						// starts fresh either way.
 						const compaction = deps.resolveCompaction?.();
 						if (compaction !== undefined) {
-							void compaction.compact(conversationId, { auto: true }).catch(() => {});
+							void compaction
+								.compact(conversationId, {
+									auto: true,
+									...(payload.modelName !== undefined ? { modelName: payload.modelName } : {}),
+								})
+								.catch(() => {});
 						}
 					}
 				});
@@ -730,7 +742,7 @@ export function createWarmService(
 }
 
 const DEFAULT_KEEP_LAST_N = 10;
-const DEFAULT_COMPACT_THRESHOLD = 350000;
+const DEFAULT_COMPACT_PERCENT = 85;
 
 const COMPACTION_SYSTEM_PROMPT =
 	"You are a conversation summarizer. Summarize the following conversation concord concisely but comprehensively. " +
@@ -772,17 +784,28 @@ export function createCompactionService(
 				return { error: "conversation too short to compact" };
 			}
 
-			// Auto mode: check threshold (default 350k if not explicitly set;
-			// 0 explicitly disables).
+			// Auto mode: check if contextSize exceeds percent of contextWindow.
 			if (opts?.auto === true) {
-				const stored = await deps.conversationStore.getCompactThreshold(conversationId);
-				const threshold = stored ?? DEFAULT_COMPACT_THRESHOLD;
-				if (threshold <= 0) return { error: "auto-compact disabled" };
+				const stored = await deps.conversationStore.getCompactPercent(conversationId);
+				const percent = stored ?? DEFAULT_COMPACT_PERCENT;
+				if (percent <= 0) return { error: "auto-compact disabled" };
 				const metrics = await deps.conversationStore.loadMetrics(conversationId);
 				const lastTurn = metrics[metrics.length - 1];
 				if (lastTurn === undefined) return { error: "no metrics" };
-				const lastInputTokens = lastTurn.usage.inputTokens + (lastTurn.usage.cacheReadTokens ?? 0);
-				if (lastInputTokens < threshold) return { error: "threshold not exceeded" };
+				const contextSize = lastTurn.contextSize;
+				if (contextSize === undefined) return { error: "no context size" };
+
+				// Resolve the model's context window.
+				const modelName = opts.modelName;
+				if (modelName === undefined || deps.resolveModelInfo === undefined) {
+					return { error: "cannot resolve model info" };
+				}
+				const info = await deps.resolveModelInfo(modelName);
+				if (info?.contextWindow === undefined) {
+					return { error: "model context window unknown" };
+				}
+				const threshold = Math.floor(info.contextWindow * (percent / 100));
+				if (contextSize < threshold) return { error: "threshold not exceeded" };
 			}
 
 			// Split: old messages to summarize + recent messages to keep.
diff --git a/packages/session-orchestrator/src/queue.test.ts b/packages/session-orchestrator/src/queue.test.ts
index 346f9c4..58de4a9 100644
--- a/packages/session-orchestrator/src/queue.test.ts
+++ b/packages/session-orchestrator/src/queue.test.ts
@@ -87,10 +87,10 @@ function createInMemoryStore(): ConversationStore & {
 		},
 		async setConversationStatus() {},
 		async replaceHistory() {},
-		async getCompactThreshold() {
+		async getCompactPercent() {
 			return null;
 		},
-		async setCompactThreshold() {},
+		async setCompactPercent() {},
 		async forkHistory() {},
 		async setCompactedFrom() {},
 	};
diff --git a/packages/transport-contract/src/index.ts b/packages/transport-contract/src/index.ts
index 5745672..02b48a0 100644
--- a/packages/transport-contract/src/index.ts
+++ b/packages/transport-contract/src/index.ts
@@ -85,11 +85,20 @@ export interface ChatRequest {
 /**
  * Response body for `GET /models` — the model catalog.
  *
- * Each entry is a model name in `<credentialName>/<model>` form: exactly the
- * string a client passes back as `ChatRequest.model`.
+ * Each entry in `models` is a model name in `<credentialName>/<model>` form:
+ * exactly the string a client passes back as `ChatRequest.model`.
+ * `modelInfo` is an optional map from the same `<credentialName>/<model>` key
+ * to model metadata (e.g. `contextWindow`). Additive — clients that only
+ * read `models` are unaffected.
  */
 export interface ModelsResponse {
 	readonly models: readonly string[];
+	readonly modelInfo?: Readonly<Record<string, ModelMetadata>>;
+}
+
+/** Per-model metadata returned alongside the model catalog. */
+export interface ModelMetadata {
+	readonly contextWindow?: number;
 }
 
 /**
@@ -583,17 +592,17 @@ export interface CompactResponse {
 }
 
 /**
- * Response for `GET /conversations/:id/compact-threshold` — the token count
+ * Response for `GET /conversations/:id/compact-percent` — the token count
  * at which automatic compaction triggers (0 = manual only).
  */
-export interface CompactThresholdResponse {
+export interface CompactPercentResponse {
 	readonly conversationId: string;
 	readonly threshold: number;
 }
 
 /**
- * Request body for `PUT /conversations/:id/compact-threshold`.
+ * Request body for `PUT /conversations/:id/compact-percent`.
  */
-export interface SetCompactThresholdRequest {
+export interface SetCompactPercentRequest {
 	readonly threshold: number;
 }
diff --git a/packages/transport-http/src/app.test.ts b/packages/transport-http/src/app.test.ts
index 674fd94..11f74f5 100644
--- a/packages/transport-http/src/app.test.ts
+++ b/packages/transport-http/src/app.test.ts
@@ -139,10 +139,10 @@ function createFakeConversationStore(
 		},
 		async setConversationStatus() {},
 		async replaceHistory() {},
-		async getCompactThreshold() {
+		async getCompactPercent() {
 			return null;
 		},
-		async setCompactThreshold() {},
+		async setCompactPercent() {},
 		async forkHistory() {},
 		async setCompactedFrom() {},
 	};
@@ -241,6 +241,9 @@ function createFakeCredentialStore(models: string[]): CredentialStore {
 		resolve() {
 			return undefined;
 		},
+		async getModelInfo() {
+			return undefined;
+		},
 		async listCatalog() {
 			return models;
 		},
@@ -252,6 +255,9 @@ function createThrowingCredentialStore(error: Error): CredentialStore {
 		resolve() {
 			return undefined;
 		},
+		async getModelInfo() {
+			return undefined;
+		},
 		async listCatalog() {
 			throw error;
 		},
@@ -876,10 +882,10 @@ describe("GET /conversations/:id", () => {
 				},
 				async setConversationStatus() {},
 				async replaceHistory() {},
-				async getCompactThreshold() {
+				async getCompactPercent() {
 					return null;
 				},
-				async setCompactThreshold() {},
+				async setCompactPercent() {},
 				async forkHistory() {},
 				async setCompactedFrom() {},
 			};
@@ -951,10 +957,10 @@ describe("GET /conversations/:id", () => {
 			},
 			async setConversationStatus() {},
 			async replaceHistory() {},
-			async getCompactThreshold() {
+			async getCompactPercent() {
 				return null;
 			},
-			async setCompactThreshold() {},
+			async setCompactPercent() {},
 			async forkHistory() {},
 			async setCompactedFrom() {},
 		};
@@ -1095,10 +1101,10 @@ describe("GET /conversations/:id/metrics", () => {
 			},
 			async setConversationStatus() {},
 			async replaceHistory() {},
-			async getCompactThreshold() {
+			async getCompactPercent() {
 				return null;
 			},
-			async setCompactThreshold() {},
+			async setCompactPercent() {},
 			async forkHistory() {},
 			async setCompactedFrom() {},
 		};
@@ -2072,10 +2078,10 @@ describe("PUT /conversations/:id/reasoning-effort", () => {
 			},
 			async setConversationStatus() {},
 			async replaceHistory() {},
-			async getCompactThreshold() {
+			async getCompactPercent() {
 				return null;
 			},
-			async setCompactThreshold() {},
+			async setCompactPercent() {},
 			async forkHistory() {},
 			async setCompactedFrom() {},
 		};
diff --git a/packages/transport-http/src/app.ts b/packages/transport-http/src/app.ts
index a2ee18c..f917846 100644
--- a/packages/transport-http/src/app.ts
+++ b/packages/transport-http/src/app.ts
@@ -1,8 +1,8 @@
 import type { AgentEvent, HostAPI, Logger } from "@dispatch/kernel";
 import type {
 	CloseConversationResponse,
+	CompactPercentResponse,
 	CompactResponse,
-	CompactThresholdResponse,
 	ConversationHistoryResponse,
 	ConversationListResponse,
 	ConversationMetricsResponse,
@@ -14,7 +14,7 @@ import type {
 	OpenConversationResponse,
 	QueueResponse,
 	ReasoningEffortResponse,
-	SetCompactThresholdRequest,
+	SetCompactPercentRequest,
 	ThroughputResponse,
 	TitleResponse,
 	WarmResponse,
@@ -233,7 +233,17 @@ export function createApp(opts: CreateServerOptions): Hono {
 	app.get("/models", async (c) => {
 		try {
 			const models = await opts.credentialStore.listCatalog();
-			const body: ModelsResponse = { models };
+			const modelInfo: Record<string, { contextWindow?: number }> = {};
+			for (const modelName of models) {
+				const info = await opts.credentialStore.getModelInfo(modelName);
+				if (info?.contextWindow !== undefined) {
+					modelInfo[modelName] = { contextWindow: info.contextWindow };
+				}
+			}
+			const body: ModelsResponse = {
+				models,
+				...(Object.keys(modelInfo).length > 0 ? { modelInfo } : {}),
+			};
 			return c.json(body, 200);
 		} catch (err) {
 			log.error("models: failed to retrieve catalog", { err });
@@ -739,14 +749,14 @@ export function createApp(opts: CreateServerOptions): Hono {
 		return c.json(response, 200);
 	});
 
-	app.get("/conversations/:id/compact-threshold", async (c) => {
+	app.get("/conversations/:id/compact-percent", async (c) => {
 		const conversationId = c.req.param("id");
-		const threshold = (await opts.conversationStore.getCompactThreshold(conversationId)) ?? 0;
-		const response: CompactThresholdResponse = { conversationId, threshold };
+		const threshold = (await opts.conversationStore.getCompactPercent(conversationId)) ?? 0;
+		const response: CompactPercentResponse = { conversationId, threshold };
 		return c.json(response, 200);
 	});
 
-	app.put("/conversations/:id/compact-threshold", async (c) => {
+	app.put("/conversations/:id/compact-percent", async (c) => {
 		const conversationId = c.req.param("id");
 		let body: unknown;
 		try {
@@ -754,7 +764,7 @@ export function createApp(opts: CreateServerOptions): Hono {
 		} catch {
 			return c.json({ error: "Invalid JSON body" }, 400);
 		}
-		const parsed = body as SetCompactThresholdRequest;
+		const parsed = body as SetCompactPercentRequest;
 		if (
 			typeof parsed.threshold !== "number" ||
 			!Number.isFinite(parsed.threshold) ||
@@ -763,9 +773,9 @@ export function createApp(opts: CreateServerOptions): Hono {
 			return c.json({ error: "threshold must be a non-negative number" }, 400);
 		}
 		const threshold = Math.floor(parsed.threshold);
-		await opts.conversationStore.setCompactThreshold(conversationId, threshold);
-		log.info("conversations: compact-threshold set", { conversationId, threshold });
-		const response: CompactThresholdResponse = { conversationId, threshold };
+		await opts.conversationStore.setCompactPercent(conversationId, threshold);
+		log.info("conversations: compact-percent set", { conversationId, threshold });
+		const response: CompactPercentResponse = { conversationId, threshold };
 		return c.json(response, 200);
 	});
 
diff --git a/packages/transport-http/src/extension.ts b/packages/transport-http/src/extension.ts
index f7af615..36314e6 100644
--- a/packages/transport-http/src/extension.ts
+++ b/packages/transport-http/src/extension.ts
@@ -32,7 +32,7 @@ export const manifest: Manifest = {
 			"/conversations/:id",
 			"/conversations/:id/close",
 			"/conversations/:id/compact",
-			"/conversations/:id/compact-threshold",
+			"/conversations/:id/compact-percent",
 			"/conversations/:id/cwd",
 			"/conversations/:id/last",
 			"/conversations/:id/lsp",
diff --git a/packages/transport-http/src/server.bun.test.ts b/packages/transport-http/src/server.bun.test.ts
index fa519af..86ae6bc 100644
--- a/packages/transport-http/src/server.bun.test.ts
+++ b/packages/transport-http/src/server.bun.test.ts
@@ -66,10 +66,10 @@ function fakeConversationStore(): ConversationStore {
 		},
 		async setConversationStatus() {},
 		async replaceHistory() {},
-		async getCompactThreshold() {
+		async getCompactPercent() {
 			return null;
 		},
-		async setCompactThreshold() {},
+		async setCompactPercent() {},
 		async forkHistory() {},
 		async setCompactedFrom() {},
 	};
@@ -104,6 +104,9 @@ function fakeCredentialStore(): CredentialStore {
 		resolve() {
 			return undefined;
 		},
+		async getModelInfo() {
+			return undefined;
+		},
 		async listCatalog() {
 			return [];
 		},
diff --git a/tasks.md b/tasks.md
index ec90cf2..b2c2b87 100644
--- a/tasks.md
+++ b/tasks.md
@@ -589,8 +589,8 @@ conversation tab. Short-ID prefix resolution (4+ chars → full ID via `GET /con
      history with `[system: summary] + recent N` (ID stays the same so messaging
      is unaffected). `compactedFrom` chains backward: A → Y → X. Manual via
      `POST /conversations/:id/compact`; automatic after turn settles if
-     `compactThreshold` (default 350k) is exceeded. `GET/PUT
-     /conversations/:id/compact-threshold` for the setting. `conversation.compacted`
+     `compactThreshold` (default 85%) is exceeded. `GET/PUT
+     /conversations/:id/compact-percent` for the setting. `conversation.compacted`
      WS broadcast. CLI `dispatch compact <id>`. FE handoff:
      `frontend-compaction-handoff.md`.