diff --git a/README.md b/README.md
new file mode 100644
index 0000000..83be090
--- /dev/null
+++ b/README.md
@@ -0,0 +1,112 @@
+# Near
+Near is a P2P library which allows
+* Discovery like Android NSD, though with greater reliability and easier-to-use *NearDiscovery* API
+* Transfers among clients through an easy-to-use *NearConnect* API
+
+## NearDiscovery
+NearDiscovery takes hostname and a bunch of settings in a builder pattern for the discovery mechanism. A NearDiscovery object allows the following discovery related self-explanatory APIs:
+* ```void makeDiscoverable(String hostName);```
+* ```void makeNonDiscoverable();```
+* ```void startDiscovery();```
+* ```void stopDiscovery();```
+* ```Set<Host> getAllAvailablePeers();```
+* ```boolean isDiscoverable();```
+* ```boolean isDiscovering();```
+
+Here's how the NearDiscovery object is created
+```
+private NearDiscovery mNearDiscovery = new NearDiscovery.Builder()
+                .setContext(this)
+                .setDiscoverableTimeoutMillis(DISCOVERABLE_TIMEOUT_MILLIS)
+                .setDiscoveryTimeoutMillis(DISCOVERY_TIMEOUT_MILLIS)
+                .setDiscoverablePingIntervalMillis(DISCOVERABLE_PING_INTERVAL_MILLIS)
+                .setDiscoveryListener(getNearDiscoveryListener(), Looper.getMainLooper())
+                .build();
+```
+The Looper passed as the 2nd param of NearDiscovery.Builder.setDiscoveryListener() is for the thread on which the listener, the 1st param, should be called.
+Sample listener:
+```
+    @NonNull
+    private NearDiscovery.Listener getNearDiscoveryListener() {
+        return new NearDiscovery.Listener() {
+            @Override
+            public void onPeersUpdate(Set<Host> hosts) {
+                // Handle updates of peer list here - some peer might have got removed if it wasn't reachable anymore or some new peer might have been added
+            }
+
+            @Override
+            public void onDiscoveryTimeout() {
+                // This is called after the discovery timeout (specified in the builder) from starting discovery using the startDiscovery()
+            }
+
+            @Override
+            public void onDiscoveryFailure(Throwable e) {
+                // This is called if discovery could not be started
+            }
+
+            @Override
+            public void onDiscoverableTimeout() {
+                // This is called after the discoverable timeout (specified in the builder) from becoming discoverable by others using the makeDiscoverable()
+            }
+        };
+    }
+```
+NearDiscovery.Builder.setDiscoverablePingIntervalMillis() tells the interval at which each client broadcasts about its existence. The NearDiscovery.Listener.onPeersUpdate() gets called even if a peer is deemed stale, i.e. it's last broadcast received was more than twice the discoverable-ping-interval ago.
+
+The discovery mechanism takes place in background services which do not hold any wakelocks.
+## NearConnect
+A NearConnect object provides P2P mechanism with the following self-explanatory APIs:
+* ```long send(byte[] bytes, Host peer);```
+* ```void startReceiving();```
+* ```void stopReceiving(boolean abortCurrentTransfers);```
+* ```Set<Host> getPeers();```
+* ```boolean isReceiving();```
+
+NearConnect.startReceiving() only tells to start *listening* for any incoming transfers, similarly NearConnect.isReceiving() only tells if the client is listening for transfers or not and *not* if any data is currently being received.
+Here's how the NearConnect object is created:
+```
+        private NearConnect mNearConnect = new NearConnect.Builder()
+                .fromDiscovery(mNearDiscovery)
+                .setContext(this)
+                .setListener(getNearConnectListener(), Looper.getMainLooper())
+                .build();
+```
+The NearDiscovery object passed in NearConnect.Builder.fromDiscovery() is only to get the list of peers from. Peers can be explicitly provided as well:
+```
+        private NearConnect mNearConnect = new NearConnect.Builder()
+                .forPeers(peers) // Set<Host> peers
+                .setContext(this)
+                .setListener(getNearConnectListener(), Looper.getMainLooper()).build();
+```
+Again, the NearConnect.Builder.setListener() takes the Listener as the 1st argument and the Looper on which to call the Listener as the 2nd argument. Here's what the Listener looks like:
+```
+    @NonNull
+    private NearConnect.Listener getNearConnectListener() {
+        return new NearConnect.Listener() {
+            @Override
+            public void onReceive(byte[] bytes, final Host sender) {
+            // Process incoming data here
+            }
+
+            @Override
+            public void onSendComplete(long jobId) {
+            // jobId is the same as the return value of NearConnect.send(), an approximate epoch time of the send
+            }
+
+            @Override
+            public void onSendFailure(Throwable e, long jobId) {
+            // handle failed sends here
+            }
+
+            @Override
+            public void onStartListenFailure(Throwable e) {
+            // This tells that the NearConnect.startReceiving() didn't go through properly.
+            // Common cause would be that another instance of NearConnect is already listening and it's NearConnect.stopReceiving() needs to be called first
+            }
+        };
+    }
+```
+It's required to stop listening on a previous instance of NearConnect (using NearConnect.stopReceiving()) so as to start listening on another (by using NearConnect.startReceiving()).
+This is because the same server port is used in each instance (<- could be made configurable later if deemed necessary).
+
+on startReceiving(), and on send() partial wakelocks are held and released on stopReceiving() and on send completion/failure respectively.