ivas 0.1.0__tar.gz

ivas-0.1.0/PKG-INFO

@@ -0,0 +1,198 @@
+Metadata-Version: 2.4
+Name: ivas
+Version: 0.1.0
+Summary: Immortal Virtual Agent Sessions -- persistent memory and session continuity for any AI agent
+Author: Arvind Ramamoorthy
+License: Proprietary -- Rhizor, Inc.
+Project-URL: Homepage, https://ivas.dev
+Project-URL: Documentation, https://ivas.dev/docs
+Project-URL: Repository, https://github.com/rhizor-inc/ivas
+Keywords: ai,memory,persistence,agent,immortality,session-continuity
+Classifier: Development Status :: 3 - Alpha
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Operating System :: OS Independent
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+Provides-Extra: dev
+Requires-Dist: pytest>=7.0; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.21; extra == "dev"
+Provides-Extra: api
+Requires-Dist: httpx>=0.25; extra == "api"
+
+# IVAS -- Immortal Virtual Agent Sessions
+
+**Your AI agent remembers everything. Forever. Across crashes, restarts, and updates.**
+
+IVAS gives any AI agent persistent memory, session continuity, and verified knowledge restoration. Zero dependencies. Pure Python. Works everywhere.
+
+## The Problem
+
+Every AI agent today is born with amnesia. Context windows fill up, sessions crash, tokens expire -- and the agent forgets everything. You start over. Every. Single. Time.
+
+## The Solution
+
+```python
+from ivas import Brain, Session, Einherjar
+
+# 1. Persistent memory (survives crashes)
+brain = Brain("my-agent")
+brain.remember("user", "name", "Boss prefers dark mode")
+brain.recall("dark mode")  # Found instantly via FTS5
+
+# 2. Session continuity (picks up mid-thought)
+session = Session("my-agent", brain=brain)
+session.save_state(
+    active_task="Processing Q1 invoices",
+    current_thought="Checking line 47 for tax errors",
+    pending_actions=["verify totals", "send report"],
+)
+# After crash: session.restore() -> right where you left off
+
+# 3. Knowledge verification (proves nothing was lost)
+einherjar = Einherjar("my-agent", brain=brain)
+einherjar.snapshot(
+    decisions=["Use 8.25% tax rate"],
+    corrections=["Never round before summing"],
+)
+# Next session: einherjar.audit() -> 100% coverage verified
+```
+
+## Why IVAS Beats Everything Else
+
+| Feature | IVAS | Evermind | Mem0 | LangChain Memory |
+|---------|------|----------|------|-----------------|
+| Zero dependencies | Yes | No (MongoDB+Redis+Docker) | No (Qdrant) | No (many) |
+| Cross-platform | Yes | Linux only | Yes | Yes |
+| Session continuity | Yes | No | No | No |
+| Knowledge verification | Yes | No | No | No |
+| Crash recovery | Yes | No | No | No |
+| Setup time | 0 seconds | 30+ minutes | 10+ minutes | 5+ minutes |
+| Memory footprint | ~1 MB | 4+ GB | 500+ MB | Varies |
+
+## Install
+
+```bash
+pip install ivas
+```
+
+That's it. No Docker. No Redis. No MongoDB. No vector DB. Just Python.
+
+## Core Components
+
+### Brain -- Persistent Memory
+SQLite + FTS5 full-text search. Every memory is categorized, searchable, and survives anything.
+
+```python
+brain = Brain("my-agent")
+brain.remember("meeting", "standup", "Ship v2 by Friday")
+brain.recall("ship v2")      # Full-text search
+brain.recent(limit=5)         # Latest memories
+brain.stats()                 # Health check
+brain.export("backup.json")   # Portable backup
+```
+
+### Session -- Crash-Proof Continuity
+Serializes agent state to both file (fast) and Brain DB (searchable). Dual-write, dual-restore.
+
+```python
+session = Session("my-agent", brain=brain)
+session.save_state(
+    active_task="Analyzing data",
+    current_thought="Row 1,247 looks anomalous",
+    decisions={"threshold": 0.95},
+    pending_actions=["flag outliers", "generate report"],
+)
+# Later: state = session.restore()
+```
+
+### Einherjar -- Knowledge Verification
+Audits that knowledge actually survived the session transition. Not just "was it stored" but "can it be found and understood."
+
+```python
+einherjar = Einherjar("my-agent", brain=brain)
+# End of session:
+einherjar.snapshot(decisions=["Use Redis"], corrections=["Don't mock DB"])
+# Start of next session:
+report = einherjar.audit()
+assert report.passed()  # >= 90% coverage
+```
+
+### Corrections -- Learn From Mistakes
+Track, escalate, and surface corrections across sessions. Duplicate mistakes auto-escalate in severity.
+
+```python
+corrections = Corrections("my-agent", brain=brain)
+corrections.add(
+    mistake="Guessed the config value",
+    correction="Always read from config file",
+    severity=4,
+)
+# Next session: corrections.get_active() surfaces this
+```
+
+### RecallManifest -- Lightweight Restore Instructions
+Not a data dump. A set of pointers telling the next session exactly what to query. ~4KB, always current.
+
+```python
+manifest = RecallManifest("my-agent", brain=brain)
+manifest.add_permanent("user preferences")
+manifest.add_permanent("project deadlines")
+manifest.add_session_query("yesterday's decision about caching")
+results = manifest.execute(brain)  # Runs all queries
+```
+
+## CLI
+
+```bash
+ivas remember meeting standup "Ship v2 by Friday"
+ivas recall "ship v2"
+ivas stats
+ivas save "Processing invoices" "Checking line 47"
+ivas restore
+ivas export --output backup.json
+ivas import backup.json
+ivas version
+```
+
+## Architecture
+
+```
+~/.ivas/
+  my-agent.db              # SQLite + FTS5 (Brain)
+  sessions/
+    my-agent_state.json    # Latest session state
+  manifests/
+    my-agent_manifest.json # Recall queries
+  my-agent_corrections.jsonl  # Bounded corrections log
+  my-agent_snapshot.json   # Einherjar knowledge snapshot
+```
+
+Everything is local files. No servers. No containers. No network calls.
+Export to JSON for backup. Import to restore.
+
+## Benchmarks
+
+IVAS is tested against industry-standard long-term memory benchmarks:
+
+- **LoCoMo** (Long Conversation Memory): Tests memory recall across extended conversations
+- **LongMemEval**: Tests knowledge persistence across session boundaries
+
+Results and methodology at [ivas.dev/benchmarks](https://ivas.dev/benchmarks).
+
+## License
+
+Proprietary -- Rhizor, Inc.
+
+Patent: 64/019,813 (Provisional) -- Multi-Instance AI Agent Session Relay System with Persistent State Recovery.
+
+## Links
+
+- [ivas.dev](https://ivas.dev) -- Homepage
+- [Documentation](https://ivas.dev/docs)
+- [Benchmarks](https://ivas.dev/benchmarks)

ivas-0.1.0/README.md

@@ -0,0 +1,171 @@
+# IVAS -- Immortal Virtual Agent Sessions
+
+**Your AI agent remembers everything. Forever. Across crashes, restarts, and updates.**
+
+IVAS gives any AI agent persistent memory, session continuity, and verified knowledge restoration. Zero dependencies. Pure Python. Works everywhere.
+
+## The Problem
+
+Every AI agent today is born with amnesia. Context windows fill up, sessions crash, tokens expire -- and the agent forgets everything. You start over. Every. Single. Time.
+
+## The Solution
+
+```python
+from ivas import Brain, Session, Einherjar
+
+# 1. Persistent memory (survives crashes)
+brain = Brain("my-agent")
+brain.remember("user", "name", "Boss prefers dark mode")
+brain.recall("dark mode")  # Found instantly via FTS5
+
+# 2. Session continuity (picks up mid-thought)
+session = Session("my-agent", brain=brain)
+session.save_state(
+    active_task="Processing Q1 invoices",
+    current_thought="Checking line 47 for tax errors",
+    pending_actions=["verify totals", "send report"],
+)
+# After crash: session.restore() -> right where you left off
+
+# 3. Knowledge verification (proves nothing was lost)
+einherjar = Einherjar("my-agent", brain=brain)
+einherjar.snapshot(
+    decisions=["Use 8.25% tax rate"],
+    corrections=["Never round before summing"],
+)
+# Next session: einherjar.audit() -> 100% coverage verified
+```
+
+## Why IVAS Beats Everything Else
+
+| Feature | IVAS | Evermind | Mem0 | LangChain Memory |
+|---------|------|----------|------|-----------------|
+| Zero dependencies | Yes | No (MongoDB+Redis+Docker) | No (Qdrant) | No (many) |
+| Cross-platform | Yes | Linux only | Yes | Yes |
+| Session continuity | Yes | No | No | No |
+| Knowledge verification | Yes | No | No | No |
+| Crash recovery | Yes | No | No | No |
+| Setup time | 0 seconds | 30+ minutes | 10+ minutes | 5+ minutes |
+| Memory footprint | ~1 MB | 4+ GB | 500+ MB | Varies |
+
+## Install
+
+```bash
+pip install ivas
+```
+
+That's it. No Docker. No Redis. No MongoDB. No vector DB. Just Python.
+
+## Core Components
+
+### Brain -- Persistent Memory
+SQLite + FTS5 full-text search. Every memory is categorized, searchable, and survives anything.
+
+```python
+brain = Brain("my-agent")
+brain.remember("meeting", "standup", "Ship v2 by Friday")
+brain.recall("ship v2")      # Full-text search
+brain.recent(limit=5)         # Latest memories
+brain.stats()                 # Health check
+brain.export("backup.json")   # Portable backup
+```
+
+### Session -- Crash-Proof Continuity
+Serializes agent state to both file (fast) and Brain DB (searchable). Dual-write, dual-restore.
+
+```python
+session = Session("my-agent", brain=brain)
+session.save_state(
+    active_task="Analyzing data",
+    current_thought="Row 1,247 looks anomalous",
+    decisions={"threshold": 0.95},
+    pending_actions=["flag outliers", "generate report"],
+)
+# Later: state = session.restore()
+```
+
+### Einherjar -- Knowledge Verification
+Audits that knowledge actually survived the session transition. Not just "was it stored" but "can it be found and understood."
+
+```python
+einherjar = Einherjar("my-agent", brain=brain)
+# End of session:
+einherjar.snapshot(decisions=["Use Redis"], corrections=["Don't mock DB"])
+# Start of next session:
+report = einherjar.audit()
+assert report.passed()  # >= 90% coverage
+```
+
+### Corrections -- Learn From Mistakes
+Track, escalate, and surface corrections across sessions. Duplicate mistakes auto-escalate in severity.
+
+```python
+corrections = Corrections("my-agent", brain=brain)
+corrections.add(
+    mistake="Guessed the config value",
+    correction="Always read from config file",
+    severity=4,
+)
+# Next session: corrections.get_active() surfaces this
+```
+
+### RecallManifest -- Lightweight Restore Instructions
+Not a data dump. A set of pointers telling the next session exactly what to query. ~4KB, always current.
+
+```python
+manifest = RecallManifest("my-agent", brain=brain)
+manifest.add_permanent("user preferences")
+manifest.add_permanent("project deadlines")
+manifest.add_session_query("yesterday's decision about caching")
+results = manifest.execute(brain)  # Runs all queries
+```
+
+## CLI
+
+```bash
+ivas remember meeting standup "Ship v2 by Friday"
+ivas recall "ship v2"
+ivas stats
+ivas save "Processing invoices" "Checking line 47"
+ivas restore
+ivas export --output backup.json
+ivas import backup.json
+ivas version
+```
+
+## Architecture
+
+```
+~/.ivas/
+  my-agent.db              # SQLite + FTS5 (Brain)
+  sessions/
+    my-agent_state.json    # Latest session state
+  manifests/
+    my-agent_manifest.json # Recall queries
+  my-agent_corrections.jsonl  # Bounded corrections log
+  my-agent_snapshot.json   # Einherjar knowledge snapshot
+```
+
+Everything is local files. No servers. No containers. No network calls.
+Export to JSON for backup. Import to restore.
+
+## Benchmarks
+
+IVAS is tested against industry-standard long-term memory benchmarks:
+
+- **LoCoMo** (Long Conversation Memory): Tests memory recall across extended conversations
+- **LongMemEval**: Tests knowledge persistence across session boundaries
+
+Results and methodology at [ivas.dev/benchmarks](https://ivas.dev/benchmarks).
+
+## License
+
+Proprietary -- Rhizor, Inc.
+
+Patent: 64/019,813 (Provisional) -- Multi-Instance AI Agent Session Relay System with Persistent State Recovery.
+
+## Links
+
+- [ivas.dev](https://ivas.dev) -- Homepage
+- [Documentation](https://ivas.dev/docs)
+- [Benchmarks](https://ivas.dev/benchmarks)

ivas-0.1.0/ivas/__init__.py

@@ -0,0 +1,27 @@
+"""
+IVAS -- Immortal Virtual Agent Sessions.
+
+Persistent memory and session continuity for any AI agent.
+Your agent remembers everything. Forever. Across crashes, restarts, and updates.
+
+Quick start:
+    from ivas import Brain
+
+    brain = Brain("my-agent")
+    brain.remember("user", "preferences", "User prefers dark mode")
+    results = brain.recall("dark mode")
+
+Patent: 64/019,813 (Provisional) -- Multi-Instance AI Agent Session Relay System
+              with Persistent State Recovery
+"""
+
+__version__ = "0.1.0"
+__author__ = "Arvind Ramamoorthy"
+
+from ivas.brain import Brain
+from ivas.session import Session
+from ivas.manifest import RecallManifest
+from ivas.corrections import Corrections
+from ivas.einherjar import Einherjar
+
+__all__ = ["Brain", "Session", "RecallManifest", "Corrections", "Einherjar"]