heurchain 1.3.0 → 1.4.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +36 -0
- package/package.json +38 -38
- package/server.js +1364 -1161
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,41 @@
|
|
|
1
1
|
# HeurChain Changelog
|
|
2
2
|
|
|
3
|
+
## v1.4.1 — 2026-05-07
|
|
4
|
+
|
|
5
|
+
### Architecture
|
|
6
|
+
|
|
7
|
+
- **Briefing worker is pure memory delivery — no inference.** HeurChain is a memory substrate. The briefing worker assembles raw proc entries and session turns; it never calls an LLM. The receiving agent decides whether to surface items, query deeper via `heurchain_search`, or synthesize with its own model.
|
|
8
|
+
- `runCompressor` remains for consolidation cue compression only (background worker, separate concern).
|
|
9
|
+
|
|
10
|
+
### Changes
|
|
11
|
+
|
|
12
|
+
- Removed LLM synthesis path from `generateBriefing()`. Raw proc + session items IS the correct output, not a fallback.
|
|
13
|
+
- Default `BRIEFING_INTERVAL_MS` changed from 6 hours → 24 hours (daily cadence).
|
|
14
|
+
- Added optional webhook push: set proc key `briefing_webhook_url` for an agent and the worker will POST `{ agent, timestamp, items }` to that URL after each briefing generation (fire-and-forget, 10s timeout). Agent's orchestration layer handles the payload independently.
|
|
15
|
+
|
|
16
|
+
---
|
|
17
|
+
|
|
18
|
+
## v1.4.0 — 2026-05-07
|
|
19
|
+
|
|
20
|
+
### New: Ollama-first compressor provider chain
|
|
21
|
+
|
|
22
|
+
- `COMPRESSOR_PROVIDER`: `ollama` (default when `OLLAMA_BASE_URL` is set) | `anthropic` | `none`
|
|
23
|
+
- Ollama default model: `qwen2.5:1.5b` — runs on CPU, ~1 GB RAM, no GPU or external API required
|
|
24
|
+
- `anthropic` provider optional — requires `ANTHROPIC_API_KEY`
|
|
25
|
+
- `none` provider: consolidation still fires, marks keys consolidated, skips cue generation
|
|
26
|
+
- Removes hard dependency on external API for self-hosted deployments
|
|
27
|
+
|
|
28
|
+
### New: Proactive memory briefing worker
|
|
29
|
+
|
|
30
|
+
- `runProactiveBriefings()` runs on `BRIEFING_INTERVAL_MS` schedule (default 24h) plus once 30s after startup
|
|
31
|
+
- Discovers active agents by scanning `proc:agent:*` keys
|
|
32
|
+
- Assembles `"HeurChain: {item}"` reminders from proc memory and recent session buffer turns
|
|
33
|
+
- Caches result at `briefing:{agent}:latest` (Redis, 24h TTL)
|
|
34
|
+
- `/api/session-context` now returns a `briefing` field — pre-generated reminders ready at session start with zero added latency (cache hit path)
|
|
35
|
+
- Per-agent controls via proc keys: `briefing_disabled`, `briefing_interval_ms`, `briefing_instructions`, `briefing_webhook_url`
|
|
36
|
+
|
|
37
|
+
---
|
|
38
|
+
|
|
3
39
|
## v1.3.0 — 2026-05-07
|
|
4
40
|
|
|
5
41
|
### New: Frictionless REST layer
|
package/package.json
CHANGED
|
@@ -1,38 +1,38 @@
|
|
|
1
|
-
{
|
|
2
|
-
"name": "heurchain",
|
|
3
|
-
"version": "1.
|
|
4
|
-
"description": "MCP (Model Context Protocol) server for HeurChain — tiered agent memory with BM25 search, session lifecycle, procedural memory, ambient capture, and frictionless session-context injection",
|
|
5
|
-
"main": "server.js",
|
|
6
|
-
"scripts": {
|
|
7
|
-
"start": "node server.js"
|
|
8
|
-
},
|
|
9
|
-
"author": "Peter Medina <peter@carlssoncreative.com>",
|
|
10
|
-
"files": [
|
|
11
|
-
"server.js",
|
|
12
|
-
"AGENT_CONFIG.json",
|
|
13
|
-
"CHANGELOG.md"
|
|
14
|
-
],
|
|
15
|
-
"keywords": [
|
|
16
|
-
"mcp",
|
|
17
|
-
"heurchain",
|
|
18
|
-
"agent-memory",
|
|
19
|
-
"llm",
|
|
20
|
-
"ai",
|
|
21
|
-
"model-context-protocol",
|
|
22
|
-
"bm25",
|
|
23
|
-
"redis",
|
|
24
|
-
"claude",
|
|
25
|
-
"anthropic"
|
|
26
|
-
],
|
|
27
|
-
"license": "MIT",
|
|
28
|
-
"engines": {
|
|
29
|
-
"node": ">=22"
|
|
30
|
-
},
|
|
31
|
-
"dependencies": {
|
|
32
|
-
"express": "^5.1.0",
|
|
33
|
-
"redis": "^5.0.0",
|
|
34
|
-
"cors": "^2.8.5",
|
|
35
|
-
"@modelcontextprotocol/sdk": "^1.17.0",
|
|
36
|
-
"zod": "^3.25.0"
|
|
37
|
-
}
|
|
38
|
-
}
|
|
1
|
+
{
|
|
2
|
+
"name": "heurchain",
|
|
3
|
+
"version": "1.4.1",
|
|
4
|
+
"description": "MCP (Model Context Protocol) server for HeurChain — tiered agent memory with BM25 search, session lifecycle, procedural memory, ambient capture, and frictionless session-context injection",
|
|
5
|
+
"main": "server.js",
|
|
6
|
+
"scripts": {
|
|
7
|
+
"start": "node server.js"
|
|
8
|
+
},
|
|
9
|
+
"author": "Peter Medina <peter@carlssoncreative.com>",
|
|
10
|
+
"files": [
|
|
11
|
+
"server.js",
|
|
12
|
+
"AGENT_CONFIG.json",
|
|
13
|
+
"CHANGELOG.md"
|
|
14
|
+
],
|
|
15
|
+
"keywords": [
|
|
16
|
+
"mcp",
|
|
17
|
+
"heurchain",
|
|
18
|
+
"agent-memory",
|
|
19
|
+
"llm",
|
|
20
|
+
"ai",
|
|
21
|
+
"model-context-protocol",
|
|
22
|
+
"bm25",
|
|
23
|
+
"redis",
|
|
24
|
+
"claude",
|
|
25
|
+
"anthropic"
|
|
26
|
+
],
|
|
27
|
+
"license": "MIT",
|
|
28
|
+
"engines": {
|
|
29
|
+
"node": ">=22"
|
|
30
|
+
},
|
|
31
|
+
"dependencies": {
|
|
32
|
+
"express": "^5.1.0",
|
|
33
|
+
"redis": "^5.0.0",
|
|
34
|
+
"cors": "^2.8.5",
|
|
35
|
+
"@modelcontextprotocol/sdk": "^1.17.0",
|
|
36
|
+
"zod": "^3.25.0"
|
|
37
|
+
}
|
|
38
|
+
}
|